diff --git a/Code/MonoMutliViewClassifiers/ExecClassif.py b/Code/MonoMutliViewClassifiers/ExecClassif.py index 78df72e5a9548786698a24ae44ad08115681ed61..34ab48ffc5028b79c395c61e861dbcda9730f4b4 100644 --- a/Code/MonoMutliViewClassifiers/ExecClassif.py +++ b/Code/MonoMutliViewClassifiers/ExecClassif.py @@ -329,6 +329,8 @@ try: argumentDictionaries["Multiview"].append(arguments) except: pass +# bestClassifiers = ["DecisionTree", "DecisionTree", "DecisionTree", "DecisionTree"] + try: if benchmark["Multiview"]["Fusion"]: try: @@ -386,7 +388,8 @@ if nbCores>1: logging.debug("Start:\t Deleting datasets for multiprocessing") times = [dataBaseTime, monoviewTime, multiviewTime] +# times=[] results = (resultsMonoview, resultsMultiview) -resultAnalysis(benchmark, results, args.name) +resultAnalysis(benchmark, results, args.name, times, metrics) diff --git a/Code/MonoMutliViewClassifiers/Metrics/roc_auc_score.py b/Code/MonoMutliViewClassifiers/Metrics/roc_auc_score.py index c7f1276cfa70c6ea74387567df23e1ec7b24f295..7beda9f4ea6d744b633a6a4de275bbe580039055 100644 --- a/Code/MonoMutliViewClassifiers/Metrics/roc_auc_score.py +++ b/Code/MonoMutliViewClassifiers/Metrics/roc_auc_score.py @@ -40,5 +40,5 @@ def getConfig(**kwargs): average = kwargs["3"] except: average = "micro" - configString = "ROS AUC score using "+str(sample_weight)+" as sample_weights, "+average+" as average (higher is better)" + configString = "ROC AUC score using "+str(sample_weight)+" as sample_weights, "+average+" as average (higher is better)" return configString \ No newline at end of file diff --git a/Code/MonoMutliViewClassifiers/Multiview/ExecMultiview.py b/Code/MonoMutliViewClassifiers/Multiview/ExecMultiview.py index 3b2f8dff691fa92d51a7f2e6e9e99b42601cfb9b..89905437318f390c485a84bc030520fefe2fed7d 100644 --- a/Code/MonoMutliViewClassifiers/Multiview/ExecMultiview.py +++ b/Code/MonoMutliViewClassifiers/Multiview/ExecMultiview.py @@ -109,7 +109,7 @@ def ExecMultiview(DATASET, name, learningRate, nbFolds, nbCores, databaseType, p logging.info("\tStart:\t Fold number " + str(foldIdx + 1)) trainIndices = [index for index in range(datasetLength) if (index not in fold) and (index not in validationIndices)] DATASET_LENGTH = len(trainIndices) - classifier = classifierClass(NB_VIEW, DATASET_LENGTH, DATASET.get("labels").value, NB_CORES=nbCores, **classificationKWARGS) + classifier = classifierClass(NB_VIEW, DATASET_LENGTH, DATASET.get("labels").value[trainIndices], NB_CORES=nbCores, **classificationKWARGS) classifier.fit_hdf5(DATASET, trainIndices=trainIndices) kFoldClassifier.append(classifier) diff --git a/Code/MonoMutliViewClassifiers/Multiview/Fusion/Fusion.py b/Code/MonoMutliViewClassifiers/Multiview/Fusion/Fusion.py index 4f1e43a4b22227e90f43ff596a989b9ef92d25aa..dc4360cc563ad7533fb611c6f363cdfc9d9322d2 100644 --- a/Code/MonoMutliViewClassifiers/Multiview/Fusion/Fusion.py +++ b/Code/MonoMutliViewClassifiers/Multiview/Fusion/Fusion.py @@ -44,7 +44,6 @@ def gridSearch_hdf5(DATASET, classificationKWARGS, learningIndices, metric=None, nIter=nIter)) logging.debug("\tDone:\t Random search for "+classifierName) classificationKWARGS["classifiersConfigs"] = bestSettings - print bestSettings fusionMethodConfig = fusionMethodModule.gridSearch(DATASET, classificationKWARGS, learningIndices, nIter=nIter) return bestSettings, fusionMethodConfig diff --git a/Code/MonoMutliViewClassifiers/Multiview/Fusion/analyzeResults.py b/Code/MonoMutliViewClassifiers/Multiview/Fusion/analyzeResults.py index 79d9f9b3be7e8b9bd2325307d70422285b2bae1b..557b408adf88898190f71e293de3c6decb530674 100644 --- a/Code/MonoMutliViewClassifiers/Multiview/Fusion/analyzeResults.py +++ b/Code/MonoMutliViewClassifiers/Multiview/Fusion/analyzeResults.py @@ -15,11 +15,11 @@ __author__ = "Baptiste Bauvin" __status__ = "Prototype" # Production, Development, Prototype - def error(testLabels, computedLabels): error = sum(map(operator.ne, computedLabels, testLabels)) return float(error) * 100 / len(computedLabels) + def getMetricScore(metric, y_train, y_train_pred, y_test, y_test_pred): metricModule = getattr(Metrics, metric[0]) if metric[1]!=None: @@ -32,6 +32,7 @@ def getMetricScore(metric, y_train, y_train_pred, y_test, y_test_pred): metricScoreString += "\n" return metricScoreString + def getTotalMetricScores(metric, kFoldPredictedTrainLabels, kFoldPredictedTestLabels, kFoldPredictedValidationLabels, DATASET, validationIndices, kFolds): labels = DATASET.get("labels").value @@ -40,7 +41,7 @@ def getTotalMetricScores(metric, kFoldPredictedTrainLabels, kFoldPredictedTestLa metricKWARGS = dict((index, metricConfig) for index, metricConfig in enumerate(metric[1])) else: metricKWARGS = {} - trainScore = np.mean(np.array([metricModule.score([label for index, label in enumerate(labels) if index not in fold+validationIndices], predictedLabels, **metricKWARGS) for fold, predictedLabels in zip(kFolds, kFoldPredictedTrainLabels)])) + trainScore = np.mean(np.array([metricModule.score([label for index, label in enumerate(labels) if (index not in fold) and (index not in validationIndices)], predictedLabels, **metricKWARGS) for fold, predictedLabels in zip(kFolds, kFoldPredictedTrainLabels)])) testScore = np.mean(np.array([metricModule.score(labels[fold], predictedLabels, **metricKWARGS) for fold, predictedLabels in zip(kFolds, kFoldPredictedTestLabels)])) validationScore = np.mean(np.array([metricModule.score(labels[validationIndices], predictedLabels, **metricKWARGS) for predictedLabels in kFoldPredictedValidationLabels])) return [trainScore, testScore, validationScore] @@ -55,9 +56,6 @@ def getMetricsScores(metrics, kFoldPredictedTrainLabels, kFoldPredictedTestLabel return metricsScores - - - def execute(kFoldClassifier, kFoldPredictedTrainLabels, kFoldPredictedTestLabels, kFoldPredictedValidationLabels, DATASET, classificationKWARGS, learningRate, LABELS_DICTIONARY, @@ -72,14 +70,14 @@ def execute(kFoldClassifier, kFoldPredictedTrainLabels, monoviewClassifiersConfigs = classificationKWARGS["classifiersConfigs"] fusionMethodConfig = classificationKWARGS["fusionMethodConfig"] - DATASET_LENGTH = DATASET.get("Metadata").attrs["datasetLength"]-len(validationIndices) + DATASET_LENGTH = DATASET.get("Metadata").attrs["datasetLength"] NB_CLASS = DATASET.get("Metadata").attrs["nbClass"] kFoldAccuracyOnTrain = [] kFoldAccuracyOnTest = [] kFoldAccuracyOnValidation = [] for foldIdx, fold in enumerate(kFolds): if fold != range(DATASET_LENGTH): - trainIndices = [index for index in range(DATASET_LENGTH) if index not in fold] + trainIndices = [index for index in range(DATASET_LENGTH) if (index not in fold) and (index not in validationIndices)] testLabels = CLASS_LABELS[fold] trainLabels = CLASS_LABELS[trainIndices] validationLabels = CLASS_LABELS[validationIndices] diff --git a/Code/MonoMutliViewClassifiers/ResultAnalysis.py b/Code/MonoMutliViewClassifiers/ResultAnalysis.py index 032eee4fb9b8ff998b6e785b58b7b4e851df313f..ff02817fe307d69623975821d7f52d1e4398a75b 100644 --- a/Code/MonoMutliViewClassifiers/ResultAnalysis.py +++ b/Code/MonoMutliViewClassifiers/ResultAnalysis.py @@ -8,6 +8,9 @@ import matplotlib matplotlib.use('Agg') import matplotlib.pyplot as plt +#Import own Modules +import Metrics + # Author-Info __author__ = "Baptiste Bauvin" __status__ = "Prototype" # Production, Development, Prototype @@ -25,7 +28,11 @@ def resultAnalysis(benchmark, results, name, times, metrics): fig = plt.gcf() fig.subplots_adjust(bottom=105.0, top=105.01) ax = f.add_axes([0.1, 0.1, 0.8, 0.8]) - ax.set_title(metric[0]+" on validation set for each classifier") + if metric[1]!=None: + metricKWARGS = dict((index, metricConfig) for index, metricConfig in enumerate(metric[1])) + else: + metricKWARGS = {} + ax.set_title(getattr(Metrics, metric[0]).getConfig(**metricKWARGS)+" on validation set for each classifier") ax.bar(range(nbResults), validationScores, align='center') ax.set_xticks(range(nbResults)) ax.set_xticklabels(names, rotation="vertical") diff --git a/Code/MonoMutliViewClassifiers/Results/20160906-112150-CMultiV-Benchmark-Methyl_MiRNA__RNASeq_Clinic-Fake-LOG.log b/Code/MonoMutliViewClassifiers/Results/20160906-112150-CMultiV-Benchmark-Methyl_MiRNA__RNASeq_Clinic-Fake-LOG.log new file mode 100644 index 0000000000000000000000000000000000000000..5f0724c6443589f2e3804127a17e4a946a479810 --- /dev/null +++ b/Code/MonoMutliViewClassifiers/Results/20160906-112150-CMultiV-Benchmark-Methyl_MiRNA__RNASeq_Clinic-Fake-LOG.log @@ -0,0 +1,2150 @@ +2016-09-06 11:21:50,451 DEBUG: Start: Creating 2 temporary datasets for multiprocessing +2016-09-06 11:21:50,452 WARNING: WARNING : /!\ This may use a lot of HDD storage space : 0.000124 Gbytes /!\ +2016-09-06 11:21:55,466 DEBUG: Start: Creating datasets for multiprocessing +2016-09-06 11:21:55,470 INFO: Start: Finding all available mono- & multiview algorithms +2016-09-06 11:21:55,524 DEBUG: ### Main Programm for Classification MonoView +2016-09-06 11:21:55,525 DEBUG: ### Classification - Database:Fake Feature:View0 train_size:0.7, CrossValidation k-folds:5, cores:1, algorithm : Adaboost +2016-09-06 11:21:55,525 DEBUG: Start: Determine Train/Test split +2016-09-06 11:21:55,525 DEBUG: ### Main Programm for Classification MonoView +2016-09-06 11:21:55,525 DEBUG: ### Classification - Database:Fake Feature:View0 train_size:0.7, CrossValidation k-folds:5, cores:1, algorithm : DecisionTree +2016-09-06 11:21:55,525 DEBUG: Start: Determine Train/Test split +2016-09-06 11:21:55,526 DEBUG: Info: Shape X_train:(210, 13), Length of y_train:210 +2016-09-06 11:21:55,526 DEBUG: Info: Shape X_test:(90, 13), Length of y_test:90 +2016-09-06 11:21:55,526 DEBUG: Done: Determine Train/Test split +2016-09-06 11:21:55,526 DEBUG: Info: Shape X_train:(210, 13), Length of y_train:210 +2016-09-06 11:21:55,526 DEBUG: Start: RandomSearch best settings with 1 iterations +2016-09-06 11:21:55,526 DEBUG: Info: Shape X_test:(90, 13), Length of y_test:90 +2016-09-06 11:21:55,526 DEBUG: Done: Determine Train/Test split +2016-09-06 11:21:55,526 DEBUG: Start: RandomSearch best settings with 1 iterations +2016-09-06 11:21:55,561 DEBUG: Done: RandomSearch best settings +2016-09-06 11:21:55,561 DEBUG: Start: Training +2016-09-06 11:21:55,563 DEBUG: Info: Time for Training: 0.0388250350952[s] +2016-09-06 11:21:55,563 DEBUG: Done: Training +2016-09-06 11:21:55,563 DEBUG: Start: Predicting +2016-09-06 11:21:55,565 DEBUG: Done: Predicting +2016-09-06 11:21:55,565 DEBUG: Start: Getting Results +2016-09-06 11:21:55,575 DEBUG: Done: RandomSearch best settings +2016-09-06 11:21:55,575 DEBUG: Start: Training +2016-09-06 11:21:55,579 DEBUG: Info: Time for Training: 0.0551791191101[s] +2016-09-06 11:21:55,579 DEBUG: Done: Training +2016-09-06 11:21:55,579 DEBUG: Start: Predicting +2016-09-06 11:21:55,582 DEBUG: Done: Predicting +2016-09-06 11:21:55,582 DEBUG: Start: Getting Results +2016-09-06 11:21:55,604 DEBUG: Done: Getting Results +2016-09-06 11:21:55,605 INFO: Classification on Fake database for View0 with DecisionTree + +accuracy_score on train : 1.0 +accuracy_score on test : 0.433333333333 + +Database configuration : + - Database name : Fake + - View name : View0 View shape : (300, 13) + - Learning Rate : 0.7 + - Labels used : Non, Oui + - Number of cross validation folds : 5 + +Classifier configuration : + - Decision Tree with max_depth : 15 + - Executed on 1 core(s) + - Got configuration using randomized search with 1 iterations + + + For Accuracy score using None as sample_weights (higher is better) : + - Score on train : 1.0 + - Score on test : 0.433333333333 + For F1 score using None as sample_weights, None as labels, 1 as pos_label, micro as average (higher is better) : + - Score on train : 1.0 + - Score on test : 0.413793103448 + For F-beta score using None as sample_weights, None as labels, 1 as pos_label, micro as average, 1.0 as beta (higher is better) : + - Score on train : 1.0 + - Score on test : 0.413793103448 + For Hamming loss using None as classes (lower is better) : + - Score on train : 0.0 + - Score on test : 0.566666666667 + For Jaccard similarity score using None as sample_weights (higher is better) : + - Score on train : 1.0 + - Score on test : 0.433333333333 + For Log loss using None as sample_weights, 1e-15 as eps (lower is better) : + - Score on train : nan + - Score on test : nan + For Matthews correlation coefficient (higher is better) : + - Score on train : 1.0 + - Score on test : -0.131912640639 + For Precision score using None as sample_weights, None as labels, 1 as pos_label, binary as average (higher is better) : + - Score on train : 1.0 + - Score on test : 0.391304347826 + For Recall score using None as sample_weights, None as labels, 1 as pos_label, binary as average (higher is better) : + - Score on train : 1.0 + - Score on test : 0.439024390244 + For ROS AUC score using None as sample_weights, micro as average (higher is better) : + - Score on train : 1.0 + - Score on test : 0.433797909408 + For Zero one loss using None as sample_weights (lower is better) : + - Score on train : 0.0 + - Score on test : 0.566666666667 + + + Classification took 0:00:00 +2016-09-06 11:21:55,605 INFO: Done: Result Analysis +2016-09-06 11:21:55,619 DEBUG: Done: Getting Results +2016-09-06 11:21:55,619 INFO: Classification on Fake database for View0 with Adaboost + +accuracy_score on train : 1.0 +accuracy_score on test : 0.5 + +Database configuration : + - Database name : Fake + - View name : View0 View shape : (300, 13) + - Learning Rate : 0.7 + - Labels used : Non, Oui + - Number of cross validation folds : 5 + +Classifier configuration : + - Adaboost with num_esimators : 7, base_estimators : DecisionTreeClassifier(class_weight=None, criterion='gini', max_depth=None, + max_features=None, max_leaf_nodes=None, min_samples_leaf=1, + min_samples_split=2, min_weight_fraction_leaf=0.0, + presort=False, random_state=None, splitter='best') + - Executed on 1 core(s) + - Got configuration using randomized search with 1 iterations + + + For Accuracy score using None as sample_weights (higher is better) : + - Score on train : 1.0 + - Score on test : 0.5 + For F1 score using None as sample_weights, None as labels, 1 as pos_label, micro as average (higher is better) : + - Score on train : 1.0 + - Score on test : 0.494382022472 + For F-beta score using None as sample_weights, None as labels, 1 as pos_label, micro as average, 1.0 as beta (higher is better) : + - Score on train : 1.0 + - Score on test : 0.494382022472 + For Hamming loss using None as classes (lower is better) : + - Score on train : 0.0 + - Score on test : 0.5 + For Jaccard similarity score using None as sample_weights (higher is better) : + - Score on train : 1.0 + - Score on test : 0.5 + For Log loss using None as sample_weights, 1e-15 as eps (lower is better) : + - Score on train : nan + - Score on test : nan + For Matthews correlation coefficient (higher is better) : + - Score on train : 1.0 + - Score on test : 0.00596274193664 + For Precision score using None as sample_weights, None as labels, 1 as pos_label, binary as average (higher is better) : + - Score on train : 1.0 + - Score on test : 0.458333333333 + For Recall score using None as sample_weights, None as labels, 1 as pos_label, binary as average (higher is better) : + - Score on train : 1.0 + - Score on test : 0.536585365854 + For ROS AUC score using None as sample_weights, micro as average (higher is better) : + - Score on train : 1.0 + - Score on test : 0.502986560478 + For Zero one loss using None as sample_weights (lower is better) : + - Score on train : 0.0 + - Score on test : 0.5 + + + Classification took 0:00:00 +2016-09-06 11:21:55,619 INFO: Done: Result Analysis +2016-09-06 11:21:55,774 DEBUG: ### Main Programm for Classification MonoView +2016-09-06 11:21:55,774 DEBUG: ### Main Programm for Classification MonoView +2016-09-06 11:21:55,774 DEBUG: ### Classification - Database:Fake Feature:View0 train_size:0.7, CrossValidation k-folds:5, cores:1, algorithm : KNN +2016-09-06 11:21:55,774 DEBUG: ### Classification - Database:Fake Feature:View0 train_size:0.7, CrossValidation k-folds:5, cores:1, algorithm : RandomForest +2016-09-06 11:21:55,774 DEBUG: Start: Determine Train/Test split +2016-09-06 11:21:55,774 DEBUG: Start: Determine Train/Test split +2016-09-06 11:21:55,774 DEBUG: Info: Shape X_train:(210, 13), Length of y_train:210 +2016-09-06 11:21:55,774 DEBUG: Info: Shape X_train:(210, 13), Length of y_train:210 +2016-09-06 11:21:55,775 DEBUG: Info: Shape X_test:(90, 13), Length of y_test:90 +2016-09-06 11:21:55,775 DEBUG: Info: Shape X_test:(90, 13), Length of y_test:90 +2016-09-06 11:21:55,775 DEBUG: Done: Determine Train/Test split +2016-09-06 11:21:55,775 DEBUG: Done: Determine Train/Test split +2016-09-06 11:21:55,775 DEBUG: Start: RandomSearch best settings with 1 iterations +2016-09-06 11:21:55,775 DEBUG: Start: RandomSearch best settings with 1 iterations +2016-09-06 11:21:55,806 DEBUG: Done: RandomSearch best settings +2016-09-06 11:21:55,806 DEBUG: Start: Training +2016-09-06 11:21:55,807 DEBUG: Info: Time for Training: 0.0337290763855[s] +2016-09-06 11:21:55,807 DEBUG: Done: Training +2016-09-06 11:21:55,807 DEBUG: Start: Predicting +2016-09-06 11:21:55,814 DEBUG: Done: Predicting +2016-09-06 11:21:55,814 DEBUG: Start: Getting Results +2016-09-06 11:21:55,852 DEBUG: Done: Getting Results +2016-09-06 11:21:55,852 INFO: Classification on Fake database for View0 with KNN + +accuracy_score on train : 0.557142857143 +accuracy_score on test : 0.422222222222 + +Database configuration : + - Database name : Fake + - View name : View0 View shape : (300, 13) + - Learning Rate : 0.7 + - Labels used : Non, Oui + - Number of cross validation folds : 5 + +Classifier configuration : + - K nearest Neighbors with n_neighbors: 47 + - Executed on 1 core(s) + - Got configuration using randomized search with 1 iterations + + + For Accuracy score using None as sample_weights (higher is better) : + - Score on train : 0.557142857143 + - Score on test : 0.422222222222 + For F1 score using None as sample_weights, None as labels, 1 as pos_label, micro as average (higher is better) : + - Score on train : 0.550724637681 + - Score on test : 0.315789473684 + For F-beta score using None as sample_weights, None as labels, 1 as pos_label, micro as average, 1.0 as beta (higher is better) : + - Score on train : 0.550724637681 + - Score on test : 0.315789473684 + For Hamming loss using None as classes (lower is better) : + - Score on train : 0.442857142857 + - Score on test : 0.577777777778 + For Jaccard similarity score using None as sample_weights (higher is better) : + - Score on train : 0.557142857143 + - Score on test : 0.422222222222 + For Log loss using None as sample_weights, 1e-15 as eps (lower is better) : + - Score on train : nan + - Score on test : nan + For Matthews correlation coefficient (higher is better) : + - Score on train : 0.11473701202 + - Score on test : -0.180519041032 + For Precision score using None as sample_weights, None as labels, 1 as pos_label, binary as average (higher is better) : + - Score on train : 0.564356435644 + - Score on test : 0.342857142857 + For Recall score using None as sample_weights, None as labels, 1 as pos_label, binary as average (higher is better) : + - Score on train : 0.537735849057 + - Score on test : 0.292682926829 + For ROS AUC score using None as sample_weights, micro as average (higher is better) : + - Score on train : 0.55732946299 + - Score on test : 0.411647585864 + For Zero one loss using None as sample_weights (lower is better) : + - Score on train : 0.442857142857 + - Score on test : 0.577777777778 + + + Classification took 0:00:00 +2016-09-06 11:21:55,852 INFO: Done: Result Analysis +2016-09-06 11:21:55,945 DEBUG: Done: RandomSearch best settings +2016-09-06 11:21:55,946 DEBUG: Start: Training +2016-09-06 11:21:55,964 DEBUG: Info: Time for Training: 0.190694093704[s] +2016-09-06 11:21:55,964 DEBUG: Done: Training +2016-09-06 11:21:55,964 DEBUG: Start: Predicting +2016-09-06 11:21:55,968 DEBUG: Done: Predicting +2016-09-06 11:21:55,968 DEBUG: Start: Getting Results +2016-09-06 11:21:55,996 DEBUG: Done: Getting Results +2016-09-06 11:21:55,996 INFO: Classification on Fake database for View0 with RandomForest + +accuracy_score on train : 0.957142857143 +accuracy_score on test : 0.4 + +Database configuration : + - Database name : Fake + - View name : View0 View shape : (300, 13) + - Learning Rate : 0.7 + - Labels used : Non, Oui + - Number of cross validation folds : 5 + +Classifier configuration : + - Random Forest with num_esimators : 7, max_depth : 15 + - Executed on 1 core(s) + - Got configuration using randomized search with 1 iterations + + + For Accuracy score using None as sample_weights (higher is better) : + - Score on train : 0.957142857143 + - Score on test : 0.4 + For F1 score using None as sample_weights, None as labels, 1 as pos_label, micro as average (higher is better) : + - Score on train : 0.956937799043 + - Score on test : 0.325 + For F-beta score using None as sample_weights, None as labels, 1 as pos_label, micro as average, 1.0 as beta (higher is better) : + - Score on train : 0.956937799043 + - Score on test : 0.325 + For Hamming loss using None as classes (lower is better) : + - Score on train : 0.0428571428571 + - Score on test : 0.6 + For Jaccard similarity score using None as sample_weights (higher is better) : + - Score on train : 0.957142857143 + - Score on test : 0.4 + For Log loss using None as sample_weights, 1e-15 as eps (lower is better) : + - Score on train : nan + - Score on test : nan + For Matthews correlation coefficient (higher is better) : + - Score on train : 0.914674537841 + - Score on test : -0.214609988978 + For Precision score using None as sample_weights, None as labels, 1 as pos_label, binary as average (higher is better) : + - Score on train : 0.970873786408 + - Score on test : 0.333333333333 + For Recall score using None as sample_weights, None as labels, 1 as pos_label, binary as average (higher is better) : + - Score on train : 0.943396226415 + - Score on test : 0.317073170732 + For ROS AUC score using None as sample_weights, micro as average (higher is better) : + - Score on train : 0.957275036284 + - Score on test : 0.393230462917 + For Zero one loss using None as sample_weights (lower is better) : + - Score on train : 0.0428571428571 + - Score on test : 0.6 + + + Classification took 0:00:00 +2016-09-06 11:21:55,997 INFO: Done: Result Analysis +2016-09-06 11:21:56,124 DEBUG: ### Main Programm for Classification MonoView +2016-09-06 11:21:56,125 DEBUG: ### Classification - Database:Fake Feature:View0 train_size:0.7, CrossValidation k-folds:5, cores:1, algorithm : SGD +2016-09-06 11:21:56,125 DEBUG: ### Main Programm for Classification MonoView +2016-09-06 11:21:56,125 DEBUG: Start: Determine Train/Test split +2016-09-06 11:21:56,125 DEBUG: ### Classification - Database:Fake Feature:View0 train_size:0.7, CrossValidation k-folds:5, cores:1, algorithm : SVMLinear +2016-09-06 11:21:56,125 DEBUG: Start: Determine Train/Test split +2016-09-06 11:21:56,125 DEBUG: Info: Shape X_train:(210, 13), Length of y_train:210 +2016-09-06 11:21:56,125 DEBUG: Info: Shape X_train:(210, 13), Length of y_train:210 +2016-09-06 11:21:56,126 DEBUG: Info: Shape X_test:(90, 13), Length of y_test:90 +2016-09-06 11:21:56,126 DEBUG: Done: Determine Train/Test split +2016-09-06 11:21:56,126 DEBUG: Info: Shape X_test:(90, 13), Length of y_test:90 +2016-09-06 11:21:56,126 DEBUG: Done: Determine Train/Test split +2016-09-06 11:21:56,126 DEBUG: Start: RandomSearch best settings with 1 iterations +2016-09-06 11:21:56,126 DEBUG: Start: RandomSearch best settings with 1 iterations +2016-09-06 11:21:56,173 DEBUG: Done: RandomSearch best settings +2016-09-06 11:21:56,173 DEBUG: Start: Training +2016-09-06 11:21:56,174 DEBUG: Info: Time for Training: 0.0500919818878[s] +2016-09-06 11:21:56,174 DEBUG: Done: Training +2016-09-06 11:21:56,174 DEBUG: Start: Predicting +2016-09-06 11:21:56,181 DEBUG: Done: RandomSearch best settings +2016-09-06 11:21:56,181 DEBUG: Start: Training +2016-09-06 11:21:56,185 DEBUG: Done: Predicting +2016-09-06 11:21:56,185 DEBUG: Start: Getting Results +2016-09-06 11:21:56,202 DEBUG: Info: Time for Training: 0.0777740478516[s] +2016-09-06 11:21:56,202 DEBUG: Done: Training +2016-09-06 11:21:56,202 DEBUG: Start: Predicting +2016-09-06 11:21:56,206 DEBUG: Done: Predicting +2016-09-06 11:21:56,206 DEBUG: Start: Getting Results +2016-09-06 11:21:56,213 DEBUG: Done: Getting Results +2016-09-06 11:21:56,213 INFO: Classification on Fake database for View0 with SGD + +accuracy_score on train : 0.604761904762 +accuracy_score on test : 0.433333333333 + +Database configuration : + - Database name : Fake + - View name : View0 View shape : (300, 13) + - Learning Rate : 0.7 + - Labels used : Non, Oui + - Number of cross validation folds : 5 + +Classifier configuration : + - SGDClassifier with loss : modified_huber, penalty : l2 + - Executed on 1 core(s) + - Got configuration using randomized search with 1 iterations + + + For Accuracy score using None as sample_weights (higher is better) : + - Score on train : 0.604761904762 + - Score on test : 0.433333333333 + For F1 score using None as sample_weights, None as labels, 1 as pos_label, micro as average (higher is better) : + - Score on train : 0.62443438914 + - Score on test : 0.43956043956 + For F-beta score using None as sample_weights, None as labels, 1 as pos_label, micro as average, 1.0 as beta (higher is better) : + - Score on train : 0.62443438914 + - Score on test : 0.43956043956 + For Hamming loss using None as classes (lower is better) : + - Score on train : 0.395238095238 + - Score on test : 0.566666666667 + For Jaccard similarity score using None as sample_weights (higher is better) : + - Score on train : 0.604761904762 + - Score on test : 0.433333333333 + For Log loss using None as sample_weights, 1e-15 as eps (lower is better) : + - Score on train : nan + - Score on test : nan + For Matthews correlation coefficient (higher is better) : + - Score on train : 0.209578877963 + - Score on test : -0.124719695673 + For Precision score using None as sample_weights, None as labels, 1 as pos_label, binary as average (higher is better) : + - Score on train : 0.6 + - Score on test : 0.4 + For Recall score using None as sample_weights, None as labels, 1 as pos_label, binary as average (higher is better) : + - Score on train : 0.650943396226 + - Score on test : 0.487804878049 + For ROS AUC score using None as sample_weights, micro as average (higher is better) : + - Score on train : 0.604317851959 + - Score on test : 0.437779990045 + For Zero one loss using None as sample_weights (lower is better) : + - Score on train : 0.395238095238 + - Score on test : 0.566666666667 + + + Classification took 0:00:00 +2016-09-06 11:21:56,214 INFO: Done: Result Analysis +2016-09-06 11:21:56,250 DEBUG: Done: Getting Results +2016-09-06 11:21:56,251 INFO: Classification on Fake database for View0 with SVMLinear + +accuracy_score on train : 0.495238095238 +accuracy_score on test : 0.444444444444 + +Database configuration : + - Database name : Fake + - View name : View0 View shape : (300, 13) + - Learning Rate : 0.7 + - Labels used : Non, Oui + - Number of cross validation folds : 5 + +Classifier configuration : + - SVM Linear with C : 2991 + - Executed on 1 core(s) + - Got configuration using randomized search with 1 iterations + + + For Accuracy score using None as sample_weights (higher is better) : + - Score on train : 0.495238095238 + - Score on test : 0.444444444444 + For F1 score using None as sample_weights, None as labels, 1 as pos_label, micro as average (higher is better) : + - Score on train : 0.504672897196 + - Score on test : 0.404761904762 + For F-beta score using None as sample_weights, None as labels, 1 as pos_label, micro as average, 1.0 as beta (higher is better) : + - Score on train : 0.504672897196 + - Score on test : 0.404761904762 + For Hamming loss using None as classes (lower is better) : + - Score on train : 0.504761904762 + - Score on test : 0.555555555556 + For Jaccard similarity score using None as sample_weights (higher is better) : + - Score on train : 0.495238095238 + - Score on test : 0.444444444444 + For Log loss using None as sample_weights, 1e-15 as eps (lower is better) : + - Score on train : nan + - Score on test : nan + For Matthews correlation coefficient (higher is better) : + - Score on train : -0.00980036362201 + - Score on test : -0.115633266975 + For Precision score using None as sample_weights, None as labels, 1 as pos_label, binary as average (higher is better) : + - Score on train : 0.5 + - Score on test : 0.395348837209 + For Recall score using None as sample_weights, None as labels, 1 as pos_label, binary as average (higher is better) : + - Score on train : 0.509433962264 + - Score on test : 0.414634146341 + For ROS AUC score using None as sample_weights, micro as average (higher is better) : + - Score on train : 0.495101596517 + - Score on test : 0.442010950722 + For Zero one loss using None as sample_weights (lower is better) : + - Score on train : 0.504761904762 + - Score on test : 0.555555555556 + + + Classification took 0:00:00 +2016-09-06 11:21:56,251 INFO: Done: Result Analysis +2016-09-06 11:21:56,375 DEBUG: ### Main Programm for Classification MonoView +2016-09-06 11:21:56,375 DEBUG: ### Main Programm for Classification MonoView +2016-09-06 11:21:56,376 DEBUG: ### Classification - Database:Fake Feature:View0 train_size:0.7, CrossValidation k-folds:5, cores:1, algorithm : SVMRBF +2016-09-06 11:21:56,376 DEBUG: ### Classification - Database:Fake Feature:View0 train_size:0.7, CrossValidation k-folds:5, cores:1, algorithm : SVMPoly +2016-09-06 11:21:56,376 DEBUG: Start: Determine Train/Test split +2016-09-06 11:21:56,376 DEBUG: Start: Determine Train/Test split +2016-09-06 11:21:56,377 DEBUG: Info: Shape X_train:(210, 13), Length of y_train:210 +2016-09-06 11:21:56,377 DEBUG: Info: Shape X_train:(210, 13), Length of y_train:210 +2016-09-06 11:21:56,377 DEBUG: Info: Shape X_test:(90, 13), Length of y_test:90 +2016-09-06 11:21:56,377 DEBUG: Info: Shape X_test:(90, 13), Length of y_test:90 +2016-09-06 11:21:56,377 DEBUG: Done: Determine Train/Test split +2016-09-06 11:21:56,377 DEBUG: Done: Determine Train/Test split +2016-09-06 11:21:56,377 DEBUG: Start: RandomSearch best settings with 1 iterations +2016-09-06 11:21:56,377 DEBUG: Start: RandomSearch best settings with 1 iterations +2016-09-06 11:21:56,447 DEBUG: Done: RandomSearch best settings +2016-09-06 11:21:56,448 DEBUG: Start: Training +2016-09-06 11:21:56,458 DEBUG: Done: RandomSearch best settings +2016-09-06 11:21:56,458 DEBUG: Start: Training +2016-09-06 11:21:56,471 DEBUG: Info: Time for Training: 0.0959920883179[s] +2016-09-06 11:21:56,471 DEBUG: Done: Training +2016-09-06 11:21:56,471 DEBUG: Start: Predicting +2016-09-06 11:21:56,479 DEBUG: Done: Predicting +2016-09-06 11:21:56,479 DEBUG: Start: Getting Results +2016-09-06 11:21:56,482 DEBUG: Info: Time for Training: 0.107470989227[s] +2016-09-06 11:21:56,482 DEBUG: Done: Training +2016-09-06 11:21:56,483 DEBUG: Start: Predicting +2016-09-06 11:21:56,488 DEBUG: Done: Predicting +2016-09-06 11:21:56,488 DEBUG: Start: Getting Results +2016-09-06 11:21:56,517 DEBUG: Done: Getting Results +2016-09-06 11:21:56,517 INFO: Classification on Fake database for View0 with SVMRBF + +accuracy_score on train : 1.0 +accuracy_score on test : 0.411111111111 + +Database configuration : + - Database name : Fake + - View name : View0 View shape : (300, 13) + - Learning Rate : 0.7 + - Labels used : Non, Oui + - Number of cross validation folds : 5 + +Classifier configuration : + - SVM Linear with C : 2991 + - Executed on 1 core(s) + - Got configuration using randomized search with 1 iterations + + + For Accuracy score using None as sample_weights (higher is better) : + - Score on train : 1.0 + - Score on test : 0.411111111111 + For F1 score using None as sample_weights, None as labels, 1 as pos_label, micro as average (higher is better) : + - Score on train : 1.0 + - Score on test : 0.345679012346 + For F-beta score using None as sample_weights, None as labels, 1 as pos_label, micro as average, 1.0 as beta (higher is better) : + - Score on train : 1.0 + - Score on test : 0.345679012346 + For Hamming loss using None as classes (lower is better) : + - Score on train : 0.0 + - Score on test : 0.588888888889 + For Jaccard similarity score using None as sample_weights (higher is better) : + - Score on train : 1.0 + - Score on test : 0.411111111111 + For Log loss using None as sample_weights, 1e-15 as eps (lower is better) : + - Score on train : nan + - Score on test : nan + For Matthews correlation coefficient (higher is better) : + - Score on train : 1.0 + - Score on test : -0.189573937423 + For Precision score using None as sample_weights, None as labels, 1 as pos_label, binary as average (higher is better) : + - Score on train : 1.0 + - Score on test : 0.35 + For Recall score using None as sample_weights, None as labels, 1 as pos_label, binary as average (higher is better) : + - Score on train : 1.0 + - Score on test : 0.341463414634 + For ROS AUC score using None as sample_weights, micro as average (higher is better) : + - Score on train : 1.0 + - Score on test : 0.405425584868 + For Zero one loss using None as sample_weights (lower is better) : + - Score on train : 0.0 + - Score on test : 0.588888888889 + + + Classification took 0:00:00 +2016-09-06 11:21:56,518 INFO: Done: Result Analysis +2016-09-06 11:21:56,536 DEBUG: Done: Getting Results +2016-09-06 11:21:56,537 INFO: Classification on Fake database for View0 with SVMPoly + +accuracy_score on train : 1.0 +accuracy_score on test : 0.5 + +Database configuration : + - Database name : Fake + - View name : View0 View shape : (300, 13) + - Learning Rate : 0.7 + - Labels used : Non, Oui + - Number of cross validation folds : 5 + +Classifier configuration : + - SVM Linear with C : 2991 + - Executed on 1 core(s) + - Got configuration using randomized search with 1 iterations + + + For Accuracy score using None as sample_weights (higher is better) : + - Score on train : 1.0 + - Score on test : 0.5 + For F1 score using None as sample_weights, None as labels, 1 as pos_label, micro as average (higher is better) : + - Score on train : 1.0 + - Score on test : 0.150943396226 + For F-beta score using None as sample_weights, None as labels, 1 as pos_label, micro as average, 1.0 as beta (higher is better) : + - Score on train : 1.0 + - Score on test : 0.150943396226 + For Hamming loss using None as classes (lower is better) : + - Score on train : 0.0 + - Score on test : 0.5 + For Jaccard similarity score using None as sample_weights (higher is better) : + - Score on train : 1.0 + - Score on test : 0.5 + For Log loss using None as sample_weights, 1e-15 as eps (lower is better) : + - Score on train : nan + - Score on test : nan + For Matthews correlation coefficient (higher is better) : + - Score on train : 1.0 + - Score on test : -0.096260040145 + For Precision score using None as sample_weights, None as labels, 1 as pos_label, binary as average (higher is better) : + - Score on train : 1.0 + - Score on test : 0.333333333333 + For Recall score using None as sample_weights, None as labels, 1 as pos_label, binary as average (higher is better) : + - Score on train : 1.0 + - Score on test : 0.0975609756098 + For ROS AUC score using None as sample_weights, micro as average (higher is better) : + - Score on train : 1.0 + - Score on test : 0.467147834744 + For Zero one loss using None as sample_weights (lower is better) : + - Score on train : 0.0 + - Score on test : 0.5 + + + Classification took 0:00:00 +2016-09-06 11:21:56,537 INFO: Done: Result Analysis +2016-09-06 11:21:56,625 DEBUG: ### Main Programm for Classification MonoView +2016-09-06 11:21:56,625 DEBUG: ### Main Programm for Classification MonoView +2016-09-06 11:21:56,625 DEBUG: ### Classification - Database:Fake Feature:View1 train_size:0.7, CrossValidation k-folds:5, cores:1, algorithm : DecisionTree +2016-09-06 11:21:56,625 DEBUG: ### Classification - Database:Fake Feature:View1 train_size:0.7, CrossValidation k-folds:5, cores:1, algorithm : Adaboost +2016-09-06 11:21:56,626 DEBUG: Start: Determine Train/Test split +2016-09-06 11:21:56,626 DEBUG: Start: Determine Train/Test split +2016-09-06 11:21:56,626 DEBUG: Info: Shape X_train:(210, 18), Length of y_train:210 +2016-09-06 11:21:56,626 DEBUG: Info: Shape X_train:(210, 18), Length of y_train:210 +2016-09-06 11:21:56,626 DEBUG: Info: Shape X_test:(90, 18), Length of y_test:90 +2016-09-06 11:21:56,626 DEBUG: Info: Shape X_test:(90, 18), Length of y_test:90 +2016-09-06 11:21:56,626 DEBUG: Done: Determine Train/Test split +2016-09-06 11:21:56,626 DEBUG: Done: Determine Train/Test split +2016-09-06 11:21:56,626 DEBUG: Start: RandomSearch best settings with 1 iterations +2016-09-06 11:21:56,626 DEBUG: Start: RandomSearch best settings with 1 iterations +2016-09-06 11:21:56,664 DEBUG: Done: RandomSearch best settings +2016-09-06 11:21:56,664 DEBUG: Start: Training +2016-09-06 11:21:56,666 DEBUG: Info: Time for Training: 0.0415709018707[s] +2016-09-06 11:21:56,666 DEBUG: Done: Training +2016-09-06 11:21:56,666 DEBUG: Start: Predicting +2016-09-06 11:21:56,669 DEBUG: Done: Predicting +2016-09-06 11:21:56,669 DEBUG: Start: Getting Results +2016-09-06 11:21:56,678 DEBUG: Done: RandomSearch best settings +2016-09-06 11:21:56,678 DEBUG: Start: Training +2016-09-06 11:21:56,684 DEBUG: Info: Time for Training: 0.059385061264[s] +2016-09-06 11:21:56,684 DEBUG: Done: Training +2016-09-06 11:21:56,684 DEBUG: Start: Predicting +2016-09-06 11:21:56,688 DEBUG: Done: Predicting +2016-09-06 11:21:56,688 DEBUG: Start: Getting Results +2016-09-06 11:21:56,718 DEBUG: Done: Getting Results +2016-09-06 11:21:56,718 INFO: Classification on Fake database for View1 with DecisionTree + +accuracy_score on train : 1.0 +accuracy_score on test : 0.444444444444 + +Database configuration : + - Database name : Fake + - View name : View1 View shape : (300, 18) + - Learning Rate : 0.7 + - Labels used : Non, Oui + - Number of cross validation folds : 5 + +Classifier configuration : + - Decision Tree with max_depth : 15 + - Executed on 1 core(s) + - Got configuration using randomized search with 1 iterations + + + For Accuracy score using None as sample_weights (higher is better) : + - Score on train : 1.0 + - Score on test : 0.444444444444 + For F1 score using None as sample_weights, None as labels, 1 as pos_label, micro as average (higher is better) : + - Score on train : 1.0 + - Score on test : 0.418604651163 + For F-beta score using None as sample_weights, None as labels, 1 as pos_label, micro as average, 1.0 as beta (higher is better) : + - Score on train : 1.0 + - Score on test : 0.418604651163 + For Hamming loss using None as classes (lower is better) : + - Score on train : 0.0 + - Score on test : 0.555555555556 + For Jaccard similarity score using None as sample_weights (higher is better) : + - Score on train : 1.0 + - Score on test : 0.444444444444 + For Log loss using None as sample_weights, 1e-15 as eps (lower is better) : + - Score on train : nan + - Score on test : nan + For Matthews correlation coefficient (higher is better) : + - Score on train : 1.0 + - Score on test : -0.111552687063 + For Precision score using None as sample_weights, None as labels, 1 as pos_label, binary as average (higher is better) : + - Score on train : 1.0 + - Score on test : 0.4 + For Recall score using None as sample_weights, None as labels, 1 as pos_label, binary as average (higher is better) : + - Score on train : 1.0 + - Score on test : 0.439024390244 + For ROS AUC score using None as sample_weights, micro as average (higher is better) : + - Score on train : 1.0 + - Score on test : 0.44400199104 + For Zero one loss using None as sample_weights (lower is better) : + - Score on train : 0.0 + - Score on test : 0.555555555556 + + + Classification took 0:00:00 +2016-09-06 11:21:56,719 INFO: Done: Result Analysis +2016-09-06 11:21:56,731 DEBUG: Done: Getting Results +2016-09-06 11:21:56,731 INFO: Classification on Fake database for View1 with Adaboost + +accuracy_score on train : 1.0 +accuracy_score on test : 0.444444444444 + +Database configuration : + - Database name : Fake + - View name : View1 View shape : (300, 18) + - Learning Rate : 0.7 + - Labels used : Non, Oui + - Number of cross validation folds : 5 + +Classifier configuration : + - Adaboost with num_esimators : 7, base_estimators : DecisionTreeClassifier(class_weight=None, criterion='gini', max_depth=None, + max_features=None, max_leaf_nodes=None, min_samples_leaf=1, + min_samples_split=2, min_weight_fraction_leaf=0.0, + presort=False, random_state=None, splitter='best') + - Executed on 1 core(s) + - Got configuration using randomized search with 1 iterations + + + For Accuracy score using None as sample_weights (higher is better) : + - Score on train : 1.0 + - Score on test : 0.444444444444 + For F1 score using None as sample_weights, None as labels, 1 as pos_label, micro as average (higher is better) : + - Score on train : 1.0 + - Score on test : 0.404761904762 + For F-beta score using None as sample_weights, None as labels, 1 as pos_label, micro as average, 1.0 as beta (higher is better) : + - Score on train : 1.0 + - Score on test : 0.404761904762 + For Hamming loss using None as classes (lower is better) : + - Score on train : 0.0 + - Score on test : 0.555555555556 + For Jaccard similarity score using None as sample_weights (higher is better) : + - Score on train : 1.0 + - Score on test : 0.444444444444 + For Log loss using None as sample_weights, 1e-15 as eps (lower is better) : + - Score on train : nan + - Score on test : nan + For Matthews correlation coefficient (higher is better) : + - Score on train : 1.0 + - Score on test : -0.115633266975 + For Precision score using None as sample_weights, None as labels, 1 as pos_label, binary as average (higher is better) : + - Score on train : 1.0 + - Score on test : 0.395348837209 + For Recall score using None as sample_weights, None as labels, 1 as pos_label, binary as average (higher is better) : + - Score on train : 1.0 + - Score on test : 0.414634146341 + For ROS AUC score using None as sample_weights, micro as average (higher is better) : + - Score on train : 1.0 + - Score on test : 0.442010950722 + For Zero one loss using None as sample_weights (lower is better) : + - Score on train : 0.0 + - Score on test : 0.555555555556 + + + Classification took 0:00:00 +2016-09-06 11:21:56,731 INFO: Done: Result Analysis +2016-09-06 11:21:56,878 DEBUG: ### Main Programm for Classification MonoView +2016-09-06 11:21:56,878 DEBUG: ### Classification - Database:Fake Feature:View1 train_size:0.7, CrossValidation k-folds:5, cores:1, algorithm : KNN +2016-09-06 11:21:56,878 DEBUG: Start: Determine Train/Test split +2016-09-06 11:21:56,879 DEBUG: Info: Shape X_train:(210, 18), Length of y_train:210 +2016-09-06 11:21:56,879 DEBUG: Info: Shape X_test:(90, 18), Length of y_test:90 +2016-09-06 11:21:56,879 DEBUG: Done: Determine Train/Test split +2016-09-06 11:21:56,879 DEBUG: Start: RandomSearch best settings with 1 iterations +2016-09-06 11:21:56,879 DEBUG: ### Main Programm for Classification MonoView +2016-09-06 11:21:56,880 DEBUG: ### Classification - Database:Fake Feature:View1 train_size:0.7, CrossValidation k-folds:5, cores:1, algorithm : RandomForest +2016-09-06 11:21:56,880 DEBUG: Start: Determine Train/Test split +2016-09-06 11:21:56,880 DEBUG: Info: Shape X_train:(210, 18), Length of y_train:210 +2016-09-06 11:21:56,881 DEBUG: Info: Shape X_test:(90, 18), Length of y_test:90 +2016-09-06 11:21:56,881 DEBUG: Done: Determine Train/Test split +2016-09-06 11:21:56,881 DEBUG: Start: RandomSearch best settings with 1 iterations +2016-09-06 11:21:56,909 DEBUG: Done: RandomSearch best settings +2016-09-06 11:21:56,909 DEBUG: Start: Training +2016-09-06 11:21:56,910 DEBUG: Info: Time for Training: 0.032821893692[s] +2016-09-06 11:21:56,910 DEBUG: Done: Training +2016-09-06 11:21:56,910 DEBUG: Start: Predicting +2016-09-06 11:21:56,917 DEBUG: Done: Predicting +2016-09-06 11:21:56,918 DEBUG: Start: Getting Results +2016-09-06 11:21:56,972 DEBUG: Done: Getting Results +2016-09-06 11:21:56,973 INFO: Classification on Fake database for View1 with KNN + +accuracy_score on train : 0.547619047619 +accuracy_score on test : 0.455555555556 + +Database configuration : + - Database name : Fake + - View name : View1 View shape : (300, 18) + - Learning Rate : 0.7 + - Labels used : Non, Oui + - Number of cross validation folds : 5 + +Classifier configuration : + - K nearest Neighbors with n_neighbors: 47 + - Executed on 1 core(s) + - Got configuration using randomized search with 1 iterations + + + For Accuracy score using None as sample_weights (higher is better) : + - Score on train : 0.547619047619 + - Score on test : 0.455555555556 + For F1 score using None as sample_weights, None as labels, 1 as pos_label, micro as average (higher is better) : + - Score on train : 0.633204633205 + - Score on test : 0.558558558559 + For F-beta score using None as sample_weights, None as labels, 1 as pos_label, micro as average, 1.0 as beta (higher is better) : + - Score on train : 0.633204633205 + - Score on test : 0.558558558559 + For Hamming loss using None as classes (lower is better) : + - Score on train : 0.452380952381 + - Score on test : 0.544444444444 + For Jaccard similarity score using None as sample_weights (higher is better) : + - Score on train : 0.547619047619 + - Score on test : 0.455555555556 + For Log loss using None as sample_weights, 1e-15 as eps (lower is better) : + - Score on train : nan + - Score on test : nan + For Matthews correlation coefficient (higher is better) : + - Score on train : 0.102191547553 + - Score on test : -0.0477019354931 + For Precision score using None as sample_weights, None as labels, 1 as pos_label, binary as average (higher is better) : + - Score on train : 0.535947712418 + - Score on test : 0.442857142857 + For Recall score using None as sample_weights, None as labels, 1 as pos_label, binary as average (higher is better) : + - Score on train : 0.77358490566 + - Score on test : 0.756097560976 + For ROS AUC score using None as sample_weights, micro as average (higher is better) : + - Score on train : 0.545446298984 + - Score on test : 0.480089596814 + For Zero one loss using None as sample_weights (lower is better) : + - Score on train : 0.452380952381 + - Score on test : 0.544444444444 + + + Classification took 0:00:00 +2016-09-06 11:21:56,973 INFO: Done: Result Analysis +2016-09-06 11:21:57,059 DEBUG: Done: RandomSearch best settings +2016-09-06 11:21:57,059 DEBUG: Start: Training +2016-09-06 11:21:57,082 DEBUG: Info: Time for Training: 0.203009128571[s] +2016-09-06 11:21:57,082 DEBUG: Done: Training +2016-09-06 11:21:57,082 DEBUG: Start: Predicting +2016-09-06 11:21:57,088 DEBUG: Done: Predicting +2016-09-06 11:21:57,088 DEBUG: Start: Getting Results +2016-09-06 11:21:57,136 DEBUG: Done: Getting Results +2016-09-06 11:21:57,136 INFO: Classification on Fake database for View1 with RandomForest + +accuracy_score on train : 0.97619047619 +accuracy_score on test : 0.522222222222 + +Database configuration : + - Database name : Fake + - View name : View1 View shape : (300, 18) + - Learning Rate : 0.7 + - Labels used : Non, Oui + - Number of cross validation folds : 5 + +Classifier configuration : + - Random Forest with num_esimators : 7, max_depth : 15 + - Executed on 1 core(s) + - Got configuration using randomized search with 1 iterations + + + For Accuracy score using None as sample_weights (higher is better) : + - Score on train : 0.97619047619 + - Score on test : 0.522222222222 + For F1 score using None as sample_weights, None as labels, 1 as pos_label, micro as average (higher is better) : + - Score on train : 0.976525821596 + - Score on test : 0.516853932584 + For F-beta score using None as sample_weights, None as labels, 1 as pos_label, micro as average, 1.0 as beta (higher is better) : + - Score on train : 0.976525821596 + - Score on test : 0.516853932584 + For Hamming loss using None as classes (lower is better) : + - Score on train : 0.0238095238095 + - Score on test : 0.477777777778 + For Jaccard similarity score using None as sample_weights (higher is better) : + - Score on train : 0.97619047619 + - Score on test : 0.522222222222 + For Log loss using None as sample_weights, 1e-15 as eps (lower is better) : + - Score on train : nan + - Score on test : nan + For Matthews correlation coefficient (higher is better) : + - Score on train : 0.952415522541 + - Score on test : 0.0506833064614 + For Precision score using None as sample_weights, None as labels, 1 as pos_label, binary as average (higher is better) : + - Score on train : 0.971962616822 + - Score on test : 0.479166666667 + For Recall score using None as sample_weights, None as labels, 1 as pos_label, binary as average (higher is better) : + - Score on train : 0.981132075472 + - Score on test : 0.560975609756 + For ROS AUC score using None as sample_weights, micro as average (higher is better) : + - Score on train : 0.976142960813 + - Score on test : 0.525385764062 + For Zero one loss using None as sample_weights (lower is better) : + - Score on train : 0.0238095238095 + - Score on test : 0.477777777778 + + + Classification took 0:00:00 +2016-09-06 11:21:57,136 INFO: Done: Result Analysis +2016-09-06 11:21:57,224 DEBUG: ### Main Programm for Classification MonoView +2016-09-06 11:21:57,224 DEBUG: ### Classification - Database:Fake Feature:View1 train_size:0.7, CrossValidation k-folds:5, cores:1, algorithm : SGD +2016-09-06 11:21:57,224 DEBUG: ### Main Programm for Classification MonoView +2016-09-06 11:21:57,225 DEBUG: Start: Determine Train/Test split +2016-09-06 11:21:57,225 DEBUG: ### Classification - Database:Fake Feature:View1 train_size:0.7, CrossValidation k-folds:5, cores:1, algorithm : SVMLinear +2016-09-06 11:21:57,225 DEBUG: Start: Determine Train/Test split +2016-09-06 11:21:57,225 DEBUG: Info: Shape X_train:(210, 18), Length of y_train:210 +2016-09-06 11:21:57,225 DEBUG: Info: Shape X_train:(210, 18), Length of y_train:210 +2016-09-06 11:21:57,225 DEBUG: Info: Shape X_test:(90, 18), Length of y_test:90 +2016-09-06 11:21:57,225 DEBUG: Done: Determine Train/Test split +2016-09-06 11:21:57,225 DEBUG: Info: Shape X_test:(90, 18), Length of y_test:90 +2016-09-06 11:21:57,226 DEBUG: Done: Determine Train/Test split +2016-09-06 11:21:57,226 DEBUG: Start: RandomSearch best settings with 1 iterations +2016-09-06 11:21:57,226 DEBUG: Start: RandomSearch best settings with 1 iterations +2016-09-06 11:21:57,271 DEBUG: Done: RandomSearch best settings +2016-09-06 11:21:57,271 DEBUG: Start: Training +2016-09-06 11:21:57,272 DEBUG: Info: Time for Training: 0.0480880737305[s] +2016-09-06 11:21:57,272 DEBUG: Done: Training +2016-09-06 11:21:57,272 DEBUG: Start: Predicting +2016-09-06 11:21:57,281 DEBUG: Done: RandomSearch best settings +2016-09-06 11:21:57,281 DEBUG: Start: Training +2016-09-06 11:21:57,288 DEBUG: Done: Predicting +2016-09-06 11:21:57,289 DEBUG: Start: Getting Results +2016-09-06 11:21:57,310 DEBUG: Info: Time for Training: 0.0865099430084[s] +2016-09-06 11:21:57,311 DEBUG: Done: Training +2016-09-06 11:21:57,311 DEBUG: Start: Predicting +2016-09-06 11:21:57,315 DEBUG: Done: Predicting +2016-09-06 11:21:57,315 DEBUG: Start: Getting Results +2016-09-06 11:21:57,316 DEBUG: Done: Getting Results +2016-09-06 11:21:57,316 INFO: Classification on Fake database for View1 with SGD + +accuracy_score on train : 0.566666666667 +accuracy_score on test : 0.522222222222 + +Database configuration : + - Database name : Fake + - View name : View1 View shape : (300, 18) + - Learning Rate : 0.7 + - Labels used : Non, Oui + - Number of cross validation folds : 5 + +Classifier configuration : + - SGDClassifier with loss : modified_huber, penalty : l2 + - Executed on 1 core(s) + - Got configuration using randomized search with 1 iterations + + + For Accuracy score using None as sample_weights (higher is better) : + - Score on train : 0.566666666667 + - Score on test : 0.522222222222 + For F1 score using None as sample_weights, None as labels, 1 as pos_label, micro as average (higher is better) : + - Score on train : 0.564593301435 + - Score on test : 0.516853932584 + For F-beta score using None as sample_weights, None as labels, 1 as pos_label, micro as average, 1.0 as beta (higher is better) : + - Score on train : 0.564593301435 + - Score on test : 0.516853932584 + For Hamming loss using None as classes (lower is better) : + - Score on train : 0.433333333333 + - Score on test : 0.477777777778 + For Jaccard similarity score using None as sample_weights (higher is better) : + - Score on train : 0.566666666667 + - Score on test : 0.522222222222 + For Log loss using None as sample_weights, 1e-15 as eps (lower is better) : + - Score on train : nan + - Score on test : nan + For Matthews correlation coefficient (higher is better) : + - Score on train : 0.133545022783 + - Score on test : 0.0506833064614 + For Precision score using None as sample_weights, None as labels, 1 as pos_label, binary as average (higher is better) : + - Score on train : 0.572815533981 + - Score on test : 0.479166666667 + For Recall score using None as sample_weights, None as labels, 1 as pos_label, binary as average (higher is better) : + - Score on train : 0.556603773585 + - Score on test : 0.560975609756 + For ROS AUC score using None as sample_weights, micro as average (higher is better) : + - Score on train : 0.566763425254 + - Score on test : 0.525385764062 + For Zero one loss using None as sample_weights (lower is better) : + - Score on train : 0.433333333333 + - Score on test : 0.477777777778 + + + Classification took 0:00:00 +2016-09-06 11:21:57,316 INFO: Done: Result Analysis +2016-09-06 11:21:57,383 DEBUG: Done: Getting Results +2016-09-06 11:21:57,383 INFO: Classification on Fake database for View1 with SVMLinear + +accuracy_score on train : 0.528571428571 +accuracy_score on test : 0.533333333333 + +Database configuration : + - Database name : Fake + - View name : View1 View shape : (300, 18) + - Learning Rate : 0.7 + - Labels used : Non, Oui + - Number of cross validation folds : 5 + +Classifier configuration : + - SVM Linear with C : 2991 + - Executed on 1 core(s) + - Got configuration using randomized search with 1 iterations + + + For Accuracy score using None as sample_weights (higher is better) : + - Score on train : 0.528571428571 + - Score on test : 0.533333333333 + For F1 score using None as sample_weights, None as labels, 1 as pos_label, micro as average (higher is better) : + - Score on train : 0.535211267606 + - Score on test : 0.475 + For F-beta score using None as sample_weights, None as labels, 1 as pos_label, micro as average, 1.0 as beta (higher is better) : + - Score on train : 0.535211267606 + - Score on test : 0.475 + For Hamming loss using None as classes (lower is better) : + - Score on train : 0.471428571429 + - Score on test : 0.466666666667 + For Jaccard similarity score using None as sample_weights (higher is better) : + - Score on train : 0.528571428571 + - Score on test : 0.533333333333 + For Log loss using None as sample_weights, 1e-15 as eps (lower is better) : + - Score on train : nan + - Score on test : nan + For Matthews correlation coefficient (higher is better) : + - Score on train : 0.0569743711331 + - Score on test : 0.0555284586866 + For Precision score using None as sample_weights, None as labels, 1 as pos_label, binary as average (higher is better) : + - Score on train : 0.532710280374 + - Score on test : 0.487179487179 + For Recall score using None as sample_weights, None as labels, 1 as pos_label, binary as average (higher is better) : + - Score on train : 0.537735849057 + - Score on test : 0.463414634146 + For ROS AUC score using None as sample_weights, micro as average (higher is better) : + - Score on train : 0.528483309144 + - Score on test : 0.52762568442 + For Zero one loss using None as sample_weights (lower is better) : + - Score on train : 0.471428571429 + - Score on test : 0.466666666667 + + + Classification took 0:00:00 +2016-09-06 11:21:57,383 INFO: Done: Result Analysis +2016-09-06 11:21:57,473 DEBUG: ### Main Programm for Classification MonoView +2016-09-06 11:21:57,473 DEBUG: ### Main Programm for Classification MonoView +2016-09-06 11:21:57,474 DEBUG: ### Classification - Database:Fake Feature:View1 train_size:0.7, CrossValidation k-folds:5, cores:1, algorithm : SVMRBF +2016-09-06 11:21:57,474 DEBUG: ### Classification - Database:Fake Feature:View1 train_size:0.7, CrossValidation k-folds:5, cores:1, algorithm : SVMPoly +2016-09-06 11:21:57,474 DEBUG: Start: Determine Train/Test split +2016-09-06 11:21:57,474 DEBUG: Start: Determine Train/Test split +2016-09-06 11:21:57,475 DEBUG: Info: Shape X_train:(210, 18), Length of y_train:210 +2016-09-06 11:21:57,475 DEBUG: Info: Shape X_train:(210, 18), Length of y_train:210 +2016-09-06 11:21:57,475 DEBUG: Info: Shape X_test:(90, 18), Length of y_test:90 +2016-09-06 11:21:57,475 DEBUG: Info: Shape X_test:(90, 18), Length of y_test:90 +2016-09-06 11:21:57,475 DEBUG: Done: Determine Train/Test split +2016-09-06 11:21:57,475 DEBUG: Done: Determine Train/Test split +2016-09-06 11:21:57,475 DEBUG: Start: RandomSearch best settings with 1 iterations +2016-09-06 11:21:57,475 DEBUG: Start: RandomSearch best settings with 1 iterations +2016-09-06 11:21:57,549 DEBUG: Done: RandomSearch best settings +2016-09-06 11:21:57,549 DEBUG: Start: Training +2016-09-06 11:21:57,561 DEBUG: Done: RandomSearch best settings +2016-09-06 11:21:57,561 DEBUG: Start: Training +2016-09-06 11:21:57,573 DEBUG: Info: Time for Training: 0.100679159164[s] +2016-09-06 11:21:57,574 DEBUG: Done: Training +2016-09-06 11:21:57,574 DEBUG: Start: Predicting +2016-09-06 11:21:57,582 DEBUG: Done: Predicting +2016-09-06 11:21:57,582 DEBUG: Start: Getting Results +2016-09-06 11:21:57,590 DEBUG: Info: Time for Training: 0.117436170578[s] +2016-09-06 11:21:57,590 DEBUG: Done: Training +2016-09-06 11:21:57,590 DEBUG: Start: Predicting +2016-09-06 11:21:57,596 DEBUG: Done: Predicting +2016-09-06 11:21:57,597 DEBUG: Start: Getting Results +2016-09-06 11:21:57,627 DEBUG: Done: Getting Results +2016-09-06 11:21:57,627 INFO: Classification on Fake database for View1 with SVMRBF + +accuracy_score on train : 1.0 +accuracy_score on test : 0.533333333333 + +Database configuration : + - Database name : Fake + - View name : View1 View shape : (300, 18) + - Learning Rate : 0.7 + - Labels used : Non, Oui + - Number of cross validation folds : 5 + +Classifier configuration : + - SVM Linear with C : 2991 + - Executed on 1 core(s) + - Got configuration using randomized search with 1 iterations + + + For Accuracy score using None as sample_weights (higher is better) : + - Score on train : 1.0 + - Score on test : 0.533333333333 + For F1 score using None as sample_weights, None as labels, 1 as pos_label, micro as average (higher is better) : + - Score on train : 1.0 + - Score on test : 0.511627906977 + For F-beta score using None as sample_weights, None as labels, 1 as pos_label, micro as average, 1.0 as beta (higher is better) : + - Score on train : 1.0 + - Score on test : 0.511627906977 + For Hamming loss using None as classes (lower is better) : + - Score on train : 0.0 + - Score on test : 0.466666666667 + For Jaccard similarity score using None as sample_weights (higher is better) : + - Score on train : 1.0 + - Score on test : 0.533333333333 + For Log loss using None as sample_weights, 1e-15 as eps (lower is better) : + - Score on train : nan + - Score on test : nan + For Matthews correlation coefficient (higher is better) : + - Score on train : 1.0 + - Score on test : 0.066931612238 + For Precision score using None as sample_weights, None as labels, 1 as pos_label, binary as average (higher is better) : + - Score on train : 1.0 + - Score on test : 0.488888888889 + For Recall score using None as sample_weights, None as labels, 1 as pos_label, binary as average (higher is better) : + - Score on train : 1.0 + - Score on test : 0.536585365854 + For ROS AUC score using None as sample_weights, micro as average (higher is better) : + - Score on train : 1.0 + - Score on test : 0.533598805376 + For Zero one loss using None as sample_weights (lower is better) : + - Score on train : 0.0 + - Score on test : 0.466666666667 + + + Classification took 0:00:00 +2016-09-06 11:21:57,627 INFO: Done: Result Analysis +2016-09-06 11:21:57,645 DEBUG: Done: Getting Results +2016-09-06 11:21:57,645 INFO: Classification on Fake database for View1 with SVMPoly + +accuracy_score on train : 1.0 +accuracy_score on test : 0.444444444444 + +Database configuration : + - Database name : Fake + - View name : View1 View shape : (300, 18) + - Learning Rate : 0.7 + - Labels used : Non, Oui + - Number of cross validation folds : 5 + +Classifier configuration : + - SVM Linear with C : 2991 + - Executed on 1 core(s) + - Got configuration using randomized search with 1 iterations + + + For Accuracy score using None as sample_weights (higher is better) : + - Score on train : 1.0 + - Score on test : 0.444444444444 + For F1 score using None as sample_weights, None as labels, 1 as pos_label, micro as average (higher is better) : + - Score on train : 1.0 + - Score on test : 0.561403508772 + For F-beta score using None as sample_weights, None as labels, 1 as pos_label, micro as average, 1.0 as beta (higher is better) : + - Score on train : 1.0 + - Score on test : 0.561403508772 + For Hamming loss using None as classes (lower is better) : + - Score on train : 0.0 + - Score on test : 0.555555555556 + For Jaccard similarity score using None as sample_weights (higher is better) : + - Score on train : 1.0 + - Score on test : 0.444444444444 + For Log loss using None as sample_weights, 1e-15 as eps (lower is better) : + - Score on train : nan + - Score on test : nan + For Matthews correlation coefficient (higher is better) : + - Score on train : 1.0 + - Score on test : -0.0715653145323 + For Precision score using None as sample_weights, None as labels, 1 as pos_label, binary as average (higher is better) : + - Score on train : 1.0 + - Score on test : 0.438356164384 + For Recall score using None as sample_weights, None as labels, 1 as pos_label, binary as average (higher is better) : + - Score on train : 1.0 + - Score on test : 0.780487804878 + For ROS AUC score using None as sample_weights, micro as average (higher is better) : + - Score on train : 1.0 + - Score on test : 0.4718765555 + For Zero one loss using None as sample_weights (lower is better) : + - Score on train : 0.0 + - Score on test : 0.555555555556 + + + Classification took 0:00:00 +2016-09-06 11:21:57,645 INFO: Done: Result Analysis +2016-09-06 11:21:57,719 DEBUG: ### Main Programm for Classification MonoView +2016-09-06 11:21:57,719 DEBUG: ### Classification - Database:Fake Feature:View2 train_size:0.7, CrossValidation k-folds:5, cores:1, algorithm : Adaboost +2016-09-06 11:21:57,719 DEBUG: ### Main Programm for Classification MonoView +2016-09-06 11:21:57,719 DEBUG: Start: Determine Train/Test split +2016-09-06 11:21:57,719 DEBUG: ### Classification - Database:Fake Feature:View2 train_size:0.7, CrossValidation k-folds:5, cores:1, algorithm : DecisionTree +2016-09-06 11:21:57,719 DEBUG: Start: Determine Train/Test split +2016-09-06 11:21:57,720 DEBUG: Info: Shape X_train:(210, 8), Length of y_train:210 +2016-09-06 11:21:57,720 DEBUG: Info: Shape X_train:(210, 8), Length of y_train:210 +2016-09-06 11:21:57,720 DEBUG: Info: Shape X_test:(90, 8), Length of y_test:90 +2016-09-06 11:21:57,720 DEBUG: Info: Shape X_test:(90, 8), Length of y_test:90 +2016-09-06 11:21:57,720 DEBUG: Done: Determine Train/Test split +2016-09-06 11:21:57,720 DEBUG: Done: Determine Train/Test split +2016-09-06 11:21:57,720 DEBUG: Start: RandomSearch best settings with 1 iterations +2016-09-06 11:21:57,720 DEBUG: Start: RandomSearch best settings with 1 iterations +2016-09-06 11:21:57,753 DEBUG: Done: RandomSearch best settings +2016-09-06 11:21:57,753 DEBUG: Start: Training +2016-09-06 11:21:57,755 DEBUG: Info: Time for Training: 0.0360541343689[s] +2016-09-06 11:21:57,755 DEBUG: Done: Training +2016-09-06 11:21:57,755 DEBUG: Start: Predicting +2016-09-06 11:21:57,757 DEBUG: Done: Predicting +2016-09-06 11:21:57,757 DEBUG: Start: Getting Results +2016-09-06 11:21:57,767 DEBUG: Done: RandomSearch best settings +2016-09-06 11:21:57,767 DEBUG: Start: Training +2016-09-06 11:21:57,771 DEBUG: Info: Time for Training: 0.052619934082[s] +2016-09-06 11:21:57,771 DEBUG: Done: Training +2016-09-06 11:21:57,771 DEBUG: Start: Predicting +2016-09-06 11:21:57,774 DEBUG: Done: Predicting +2016-09-06 11:21:57,774 DEBUG: Start: Getting Results +2016-09-06 11:21:57,802 DEBUG: Done: Getting Results +2016-09-06 11:21:57,802 INFO: Classification on Fake database for View2 with DecisionTree + +accuracy_score on train : 0.995238095238 +accuracy_score on test : 0.533333333333 + +Database configuration : + - Database name : Fake + - View name : View2 View shape : (300, 8) + - Learning Rate : 0.7 + - Labels used : Non, Oui + - Number of cross validation folds : 5 + +Classifier configuration : + - Decision Tree with max_depth : 15 + - Executed on 1 core(s) + - Got configuration using randomized search with 1 iterations + + + For Accuracy score using None as sample_weights (higher is better) : + - Score on train : 0.995238095238 + - Score on test : 0.533333333333 + For F1 score using None as sample_weights, None as labels, 1 as pos_label, micro as average (higher is better) : + - Score on train : 0.995305164319 + - Score on test : 0.511627906977 + For F-beta score using None as sample_weights, None as labels, 1 as pos_label, micro as average, 1.0 as beta (higher is better) : + - Score on train : 0.995305164319 + - Score on test : 0.511627906977 + For Hamming loss using None as classes (lower is better) : + - Score on train : 0.0047619047619 + - Score on test : 0.466666666667 + For Jaccard similarity score using None as sample_weights (higher is better) : + - Score on train : 0.995238095238 + - Score on test : 0.533333333333 + For Log loss using None as sample_weights, 1e-15 as eps (lower is better) : + - Score on train : nan + - Score on test : nan + For Matthews correlation coefficient (higher is better) : + - Score on train : 0.990519401324 + - Score on test : 0.066931612238 + For Precision score using None as sample_weights, None as labels, 1 as pos_label, binary as average (higher is better) : + - Score on train : 0.990654205607 + - Score on test : 0.488888888889 + For Recall score using None as sample_weights, None as labels, 1 as pos_label, binary as average (higher is better) : + - Score on train : 1.0 + - Score on test : 0.536585365854 + For ROS AUC score using None as sample_weights, micro as average (higher is better) : + - Score on train : 0.995192307692 + - Score on test : 0.533598805376 + For Zero one loss using None as sample_weights (lower is better) : + - Score on train : 0.0047619047619 + - Score on test : 0.466666666667 + + + Classification took 0:00:00 +2016-09-06 11:21:57,803 INFO: Done: Result Analysis +2016-09-06 11:21:57,820 DEBUG: Done: Getting Results +2016-09-06 11:21:57,820 INFO: Classification on Fake database for View2 with Adaboost + +accuracy_score on train : 1.0 +accuracy_score on test : 0.555555555556 + +Database configuration : + - Database name : Fake + - View name : View2 View shape : (300, 8) + - Learning Rate : 0.7 + - Labels used : Non, Oui + - Number of cross validation folds : 5 + +Classifier configuration : + - Adaboost with num_esimators : 7, base_estimators : DecisionTreeClassifier(class_weight=None, criterion='gini', max_depth=None, + max_features=None, max_leaf_nodes=None, min_samples_leaf=1, + min_samples_split=2, min_weight_fraction_leaf=0.0, + presort=False, random_state=None, splitter='best') + - Executed on 1 core(s) + - Got configuration using randomized search with 1 iterations + + + For Accuracy score using None as sample_weights (higher is better) : + - Score on train : 1.0 + - Score on test : 0.555555555556 + For F1 score using None as sample_weights, None as labels, 1 as pos_label, micro as average (higher is better) : + - Score on train : 1.0 + - Score on test : 0.512195121951 + For F-beta score using None as sample_weights, None as labels, 1 as pos_label, micro as average, 1.0 as beta (higher is better) : + - Score on train : 1.0 + - Score on test : 0.512195121951 + For Hamming loss using None as classes (lower is better) : + - Score on train : 0.0 + - Score on test : 0.444444444444 + For Jaccard similarity score using None as sample_weights (higher is better) : + - Score on train : 1.0 + - Score on test : 0.555555555556 + For Log loss using None as sample_weights, 1e-15 as eps (lower is better) : + - Score on train : nan + - Score on test : nan + For Matthews correlation coefficient (higher is better) : + - Score on train : 1.0 + - Score on test : 0.104031856645 + For Precision score using None as sample_weights, None as labels, 1 as pos_label, binary as average (higher is better) : + - Score on train : 1.0 + - Score on test : 0.512195121951 + For Recall score using None as sample_weights, None as labels, 1 as pos_label, binary as average (higher is better) : + - Score on train : 1.0 + - Score on test : 0.512195121951 + For ROS AUC score using None as sample_weights, micro as average (higher is better) : + - Score on train : 1.0 + - Score on test : 0.552015928323 + For Zero one loss using None as sample_weights (lower is better) : + - Score on train : 0.0 + - Score on test : 0.444444444444 + + + Classification took 0:00:00 +2016-09-06 11:21:57,820 INFO: Done: Result Analysis +2016-09-06 11:21:57,970 DEBUG: ### Main Programm for Classification MonoView +2016-09-06 11:21:57,970 DEBUG: ### Main Programm for Classification MonoView +2016-09-06 11:21:57,970 DEBUG: ### Classification - Database:Fake Feature:View2 train_size:0.7, CrossValidation k-folds:5, cores:1, algorithm : KNN +2016-09-06 11:21:57,970 DEBUG: ### Classification - Database:Fake Feature:View2 train_size:0.7, CrossValidation k-folds:5, cores:1, algorithm : RandomForest +2016-09-06 11:21:57,970 DEBUG: Start: Determine Train/Test split +2016-09-06 11:21:57,970 DEBUG: Start: Determine Train/Test split +2016-09-06 11:21:57,971 DEBUG: Info: Shape X_train:(210, 8), Length of y_train:210 +2016-09-06 11:21:57,971 DEBUG: Info: Shape X_train:(210, 8), Length of y_train:210 +2016-09-06 11:21:57,971 DEBUG: Info: Shape X_test:(90, 8), Length of y_test:90 +2016-09-06 11:21:57,971 DEBUG: Info: Shape X_test:(90, 8), Length of y_test:90 +2016-09-06 11:21:57,971 DEBUG: Done: Determine Train/Test split +2016-09-06 11:21:57,971 DEBUG: Done: Determine Train/Test split +2016-09-06 11:21:57,971 DEBUG: Start: RandomSearch best settings with 1 iterations +2016-09-06 11:21:57,971 DEBUG: Start: RandomSearch best settings with 1 iterations +2016-09-06 11:21:58,004 DEBUG: Done: RandomSearch best settings +2016-09-06 11:21:58,004 DEBUG: Start: Training +2016-09-06 11:21:58,005 DEBUG: Info: Time for Training: 0.0356760025024[s] +2016-09-06 11:21:58,005 DEBUG: Done: Training +2016-09-06 11:21:58,005 DEBUG: Start: Predicting +2016-09-06 11:21:58,013 DEBUG: Done: Predicting +2016-09-06 11:21:58,013 DEBUG: Start: Getting Results +2016-09-06 11:21:58,054 DEBUG: Done: Getting Results +2016-09-06 11:21:58,054 INFO: Classification on Fake database for View2 with KNN + +accuracy_score on train : 0.552380952381 +accuracy_score on test : 0.444444444444 + +Database configuration : + - Database name : Fake + - View name : View2 View shape : (300, 8) + - Learning Rate : 0.7 + - Labels used : Non, Oui + - Number of cross validation folds : 5 + +Classifier configuration : + - K nearest Neighbors with n_neighbors: 47 + - Executed on 1 core(s) + - Got configuration using randomized search with 1 iterations + + + For Accuracy score using None as sample_weights (higher is better) : + - Score on train : 0.552380952381 + - Score on test : 0.444444444444 + For F1 score using None as sample_weights, None as labels, 1 as pos_label, micro as average (higher is better) : + - Score on train : 0.477777777778 + - Score on test : 0.342105263158 + For F-beta score using None as sample_weights, None as labels, 1 as pos_label, micro as average, 1.0 as beta (higher is better) : + - Score on train : 0.477777777778 + - Score on test : 0.342105263158 + For Hamming loss using None as classes (lower is better) : + - Score on train : 0.447619047619 + - Score on test : 0.555555555556 + For Jaccard similarity score using None as sample_weights (higher is better) : + - Score on train : 0.552380952381 + - Score on test : 0.444444444444 + For Log loss using None as sample_weights, 1e-15 as eps (lower is better) : + - Score on train : nan + - Score on test : nan + For Matthews correlation coefficient (higher is better) : + - Score on train : 0.112597765671 + - Score on test : -0.134753650348 + For Precision score using None as sample_weights, None as labels, 1 as pos_label, binary as average (higher is better) : + - Score on train : 0.581081081081 + - Score on test : 0.371428571429 + For Recall score using None as sample_weights, None as labels, 1 as pos_label, binary as average (higher is better) : + - Score on train : 0.405660377358 + - Score on test : 0.317073170732 + For ROS AUC score using None as sample_weights, micro as average (higher is better) : + - Score on train : 0.553791727141 + - Score on test : 0.434046789447 + For Zero one loss using None as sample_weights (lower is better) : + - Score on train : 0.447619047619 + - Score on test : 0.555555555556 + + + Classification took 0:00:00 +2016-09-06 11:21:58,054 INFO: Done: Result Analysis +2016-09-06 11:21:58,134 DEBUG: Done: RandomSearch best settings +2016-09-06 11:21:58,134 DEBUG: Start: Training +2016-09-06 11:21:58,151 DEBUG: Info: Time for Training: 0.181129932404[s] +2016-09-06 11:21:58,151 DEBUG: Done: Training +2016-09-06 11:21:58,151 DEBUG: Start: Predicting +2016-09-06 11:21:58,154 DEBUG: Done: Predicting +2016-09-06 11:21:58,155 DEBUG: Start: Getting Results +2016-09-06 11:21:58,182 DEBUG: Done: Getting Results +2016-09-06 11:21:58,182 INFO: Classification on Fake database for View2 with RandomForest + +accuracy_score on train : 0.966666666667 +accuracy_score on test : 0.588888888889 + +Database configuration : + - Database name : Fake + - View name : View2 View shape : (300, 8) + - Learning Rate : 0.7 + - Labels used : Non, Oui + - Number of cross validation folds : 5 + +Classifier configuration : + - Random Forest with num_esimators : 7, max_depth : 15 + - Executed on 1 core(s) + - Got configuration using randomized search with 1 iterations + + + For Accuracy score using None as sample_weights (higher is better) : + - Score on train : 0.966666666667 + - Score on test : 0.588888888889 + For F1 score using None as sample_weights, None as labels, 1 as pos_label, micro as average (higher is better) : + - Score on train : 0.96682464455 + - Score on test : 0.53164556962 + For F-beta score using None as sample_weights, None as labels, 1 as pos_label, micro as average, 1.0 as beta (higher is better) : + - Score on train : 0.96682464455 + - Score on test : 0.53164556962 + For Hamming loss using None as classes (lower is better) : + - Score on train : 0.0333333333333 + - Score on test : 0.411111111111 + For Jaccard similarity score using None as sample_weights (higher is better) : + - Score on train : 0.966666666667 + - Score on test : 0.588888888889 + For Log loss using None as sample_weights, 1e-15 as eps (lower is better) : + - Score on train : nan + - Score on test : nan + For Matthews correlation coefficient (higher is better) : + - Score on train : 0.933375664255 + - Score on test : 0.166630556676 + For Precision score using None as sample_weights, None as labels, 1 as pos_label, binary as average (higher is better) : + - Score on train : 0.971428571429 + - Score on test : 0.552631578947 + For Recall score using None as sample_weights, None as labels, 1 as pos_label, binary as average (higher is better) : + - Score on train : 0.962264150943 + - Score on test : 0.512195121951 + For ROS AUC score using None as sample_weights, micro as average (higher is better) : + - Score on train : 0.966708998549 + - Score on test : 0.582628173221 + For Zero one loss using None as sample_weights (lower is better) : + - Score on train : 0.0333333333333 + - Score on test : 0.411111111111 + + + Classification took 0:00:00 +2016-09-06 11:21:58,182 INFO: Done: Result Analysis +2016-09-06 11:21:58,321 DEBUG: ### Main Programm for Classification MonoView +2016-09-06 11:21:58,321 DEBUG: ### Classification - Database:Fake Feature:View2 train_size:0.7, CrossValidation k-folds:5, cores:1, algorithm : SGD +2016-09-06 11:21:58,321 DEBUG: ### Main Programm for Classification MonoView +2016-09-06 11:21:58,321 DEBUG: Start: Determine Train/Test split +2016-09-06 11:21:58,321 DEBUG: ### Classification - Database:Fake Feature:View2 train_size:0.7, CrossValidation k-folds:5, cores:1, algorithm : SVMLinear +2016-09-06 11:21:58,322 DEBUG: Start: Determine Train/Test split +2016-09-06 11:21:58,322 DEBUG: Info: Shape X_train:(210, 8), Length of y_train:210 +2016-09-06 11:21:58,322 DEBUG: Info: Shape X_train:(210, 8), Length of y_train:210 +2016-09-06 11:21:58,322 DEBUG: Info: Shape X_test:(90, 8), Length of y_test:90 +2016-09-06 11:21:58,322 DEBUG: Info: Shape X_test:(90, 8), Length of y_test:90 +2016-09-06 11:21:58,323 DEBUG: Done: Determine Train/Test split +2016-09-06 11:21:58,323 DEBUG: Done: Determine Train/Test split +2016-09-06 11:21:58,323 DEBUG: Start: RandomSearch best settings with 1 iterations +2016-09-06 11:21:58,323 DEBUG: Start: RandomSearch best settings with 1 iterations +2016-09-06 11:21:58,392 DEBUG: Done: RandomSearch best settings +2016-09-06 11:21:58,392 DEBUG: Start: Training +2016-09-06 11:21:58,393 DEBUG: Info: Time for Training: 0.0729160308838[s] +2016-09-06 11:21:58,393 DEBUG: Done: Training +2016-09-06 11:21:58,393 DEBUG: Start: Predicting +2016-09-06 11:21:58,396 DEBUG: Done: RandomSearch best settings +2016-09-06 11:21:58,396 DEBUG: Start: Training +2016-09-06 11:21:58,406 DEBUG: Done: Predicting +2016-09-06 11:21:58,406 DEBUG: Start: Getting Results +2016-09-06 11:21:58,415 DEBUG: Info: Time for Training: 0.095123052597[s] +2016-09-06 11:21:58,416 DEBUG: Done: Training +2016-09-06 11:21:58,416 DEBUG: Start: Predicting +2016-09-06 11:21:58,419 DEBUG: Done: Predicting +2016-09-06 11:21:58,419 DEBUG: Start: Getting Results +2016-09-06 11:21:58,431 DEBUG: Done: Getting Results +2016-09-06 11:21:58,431 INFO: Classification on Fake database for View2 with SGD + +accuracy_score on train : 0.561904761905 +accuracy_score on test : 0.5 + +Database configuration : + - Database name : Fake + - View name : View2 View shape : (300, 8) + - Learning Rate : 0.7 + - Labels used : Non, Oui + - Number of cross validation folds : 5 + +Classifier configuration : + - SGDClassifier with loss : modified_huber, penalty : l2 + - Executed on 1 core(s) + - Got configuration using randomized search with 1 iterations + + + For Accuracy score using None as sample_weights (higher is better) : + - Score on train : 0.561904761905 + - Score on test : 0.5 + For F1 score using None as sample_weights, None as labels, 1 as pos_label, micro as average (higher is better) : + - Score on train : 0.544554455446 + - Score on test : 0.470588235294 + For F-beta score using None as sample_weights, None as labels, 1 as pos_label, micro as average, 1.0 as beta (higher is better) : + - Score on train : 0.544554455446 + - Score on test : 0.470588235294 + For Hamming loss using None as classes (lower is better) : + - Score on train : 0.438095238095 + - Score on test : 0.5 + For Jaccard similarity score using None as sample_weights (higher is better) : + - Score on train : 0.561904761905 + - Score on test : 0.5 + For Log loss using None as sample_weights, 1e-15 as eps (lower is better) : + - Score on train : nan + - Score on test : nan + For Matthews correlation coefficient (higher is better) : + - Score on train : 0.125091870983 + - Score on test : -0.00198364873142 + For Precision score using None as sample_weights, None as labels, 1 as pos_label, binary as average (higher is better) : + - Score on train : 0.572916666667 + - Score on test : 0.454545454545 + For Recall score using None as sample_weights, None as labels, 1 as pos_label, binary as average (higher is better) : + - Score on train : 0.518867924528 + - Score on test : 0.487804878049 + For ROS AUC score using None as sample_weights, micro as average (higher is better) : + - Score on train : 0.562318577649 + - Score on test : 0.499004479841 + For Zero one loss using None as sample_weights (lower is better) : + - Score on train : 0.438095238095 + - Score on test : 0.5 + + + Classification took 0:00:00 +2016-09-06 11:21:58,431 INFO: Done: Result Analysis +2016-09-06 11:21:58,448 DEBUG: Done: Getting Results +2016-09-06 11:21:58,448 INFO: Classification on Fake database for View2 with SVMLinear + +accuracy_score on train : 0.495238095238 +accuracy_score on test : 0.633333333333 + +Database configuration : + - Database name : Fake + - View name : View2 View shape : (300, 8) + - Learning Rate : 0.7 + - Labels used : Non, Oui + - Number of cross validation folds : 5 + +Classifier configuration : + - SVM Linear with C : 2991 + - Executed on 1 core(s) + - Got configuration using randomized search with 1 iterations + + + For Accuracy score using None as sample_weights (higher is better) : + - Score on train : 0.495238095238 + - Score on test : 0.633333333333 + For F1 score using None as sample_weights, None as labels, 1 as pos_label, micro as average (higher is better) : + - Score on train : 0.51376146789 + - Score on test : 0.637362637363 + For F-beta score using None as sample_weights, None as labels, 1 as pos_label, micro as average, 1.0 as beta (higher is better) : + - Score on train : 0.51376146789 + - Score on test : 0.637362637363 + For Hamming loss using None as classes (lower is better) : + - Score on train : 0.504761904762 + - Score on test : 0.366666666667 + For Jaccard similarity score using None as sample_weights (higher is better) : + - Score on train : 0.495238095238 + - Score on test : 0.633333333333 + For Log loss using None as sample_weights, 1e-15 as eps (lower is better) : + - Score on train : nan + - Score on test : nan + For Matthews correlation coefficient (higher is better) : + - Score on train : -0.0101818424163 + - Score on test : 0.279372118308 + For Precision score using None as sample_weights, None as labels, 1 as pos_label, binary as average (higher is better) : + - Score on train : 0.5 + - Score on test : 0.58 + For Recall score using None as sample_weights, None as labels, 1 as pos_label, binary as average (higher is better) : + - Score on train : 0.528301886792 + - Score on test : 0.707317073171 + For ROS AUC score using None as sample_weights, micro as average (higher is better) : + - Score on train : 0.494920174165 + - Score on test : 0.6393728223 + For Zero one loss using None as sample_weights (lower is better) : + - Score on train : 0.504761904762 + - Score on test : 0.366666666667 + + + Classification took 0:00:00 +2016-09-06 11:21:58,449 INFO: Done: Result Analysis +2016-09-06 11:21:58,568 DEBUG: ### Main Programm for Classification MonoView +2016-09-06 11:21:58,568 DEBUG: ### Main Programm for Classification MonoView +2016-09-06 11:21:58,568 DEBUG: ### Classification - Database:Fake Feature:View2 train_size:0.7, CrossValidation k-folds:5, cores:1, algorithm : SVMPoly +2016-09-06 11:21:58,568 DEBUG: ### Classification - Database:Fake Feature:View2 train_size:0.7, CrossValidation k-folds:5, cores:1, algorithm : SVMRBF +2016-09-06 11:21:58,568 DEBUG: Start: Determine Train/Test split +2016-09-06 11:21:58,568 DEBUG: Start: Determine Train/Test split +2016-09-06 11:21:58,569 DEBUG: Info: Shape X_train:(210, 8), Length of y_train:210 +2016-09-06 11:21:58,569 DEBUG: Info: Shape X_train:(210, 8), Length of y_train:210 +2016-09-06 11:21:58,569 DEBUG: Info: Shape X_test:(90, 8), Length of y_test:90 +2016-09-06 11:21:58,569 DEBUG: Info: Shape X_test:(90, 8), Length of y_test:90 +2016-09-06 11:21:58,569 DEBUG: Done: Determine Train/Test split +2016-09-06 11:21:58,569 DEBUG: Done: Determine Train/Test split +2016-09-06 11:21:58,569 DEBUG: Start: RandomSearch best settings with 1 iterations +2016-09-06 11:21:58,569 DEBUG: Start: RandomSearch best settings with 1 iterations +2016-09-06 11:21:58,614 DEBUG: Done: RandomSearch best settings +2016-09-06 11:21:58,614 DEBUG: Start: Training +2016-09-06 11:21:58,619 DEBUG: Done: RandomSearch best settings +2016-09-06 11:21:58,619 DEBUG: Start: Training +2016-09-06 11:21:58,632 DEBUG: Info: Time for Training: 0.064493894577[s] +2016-09-06 11:21:58,632 DEBUG: Done: Training +2016-09-06 11:21:58,632 DEBUG: Start: Predicting +2016-09-06 11:21:58,637 DEBUG: Done: Predicting +2016-09-06 11:21:58,637 DEBUG: Start: Getting Results +2016-09-06 11:21:58,638 DEBUG: Info: Time for Training: 0.0703361034393[s] +2016-09-06 11:21:58,638 DEBUG: Done: Training +2016-09-06 11:21:58,638 DEBUG: Start: Predicting +2016-09-06 11:21:58,641 DEBUG: Done: Predicting +2016-09-06 11:21:58,641 DEBUG: Start: Getting Results +2016-09-06 11:21:58,668 DEBUG: Done: Getting Results +2016-09-06 11:21:58,668 INFO: Classification on Fake database for View2 with SVMRBF + +accuracy_score on train : 1.0 +accuracy_score on test : 0.466666666667 + +Database configuration : + - Database name : Fake + - View name : View2 View shape : (300, 8) + - Learning Rate : 0.7 + - Labels used : Non, Oui + - Number of cross validation folds : 5 + +Classifier configuration : + - SVM Linear with C : 2991 + - Executed on 1 core(s) + - Got configuration using randomized search with 1 iterations + + + For Accuracy score using None as sample_weights (higher is better) : + - Score on train : 1.0 + - Score on test : 0.466666666667 + For F1 score using None as sample_weights, None as labels, 1 as pos_label, micro as average (higher is better) : + - Score on train : 1.0 + - Score on test : 0.466666666667 + For F-beta score using None as sample_weights, None as labels, 1 as pos_label, micro as average, 1.0 as beta (higher is better) : + - Score on train : 1.0 + - Score on test : 0.466666666667 + For Hamming loss using None as classes (lower is better) : + - Score on train : 0.0 + - Score on test : 0.533333333333 + For Jaccard similarity score using None as sample_weights (higher is better) : + - Score on train : 1.0 + - Score on test : 0.466666666667 + For Log loss using None as sample_weights, 1e-15 as eps (lower is better) : + - Score on train : nan + - Score on test : nan + For Matthews correlation coefficient (higher is better) : + - Score on train : 1.0 + - Score on test : -0.0592334494774 + For Precision score using None as sample_weights, None as labels, 1 as pos_label, binary as average (higher is better) : + - Score on train : 1.0 + - Score on test : 0.428571428571 + For Recall score using None as sample_weights, None as labels, 1 as pos_label, binary as average (higher is better) : + - Score on train : 1.0 + - Score on test : 0.512195121951 + For ROS AUC score using None as sample_weights, micro as average (higher is better) : + - Score on train : 1.0 + - Score on test : 0.470383275261 + For Zero one loss using None as sample_weights (lower is better) : + - Score on train : 0.0 + - Score on test : 0.533333333333 + + + Classification took 0:00:00 +2016-09-06 11:21:58,668 INFO: Done: Result Analysis +2016-09-06 11:21:58,680 DEBUG: Done: Getting Results +2016-09-06 11:21:58,680 INFO: Classification on Fake database for View2 with SVMPoly + +accuracy_score on train : 0.995238095238 +accuracy_score on test : 0.444444444444 + +Database configuration : + - Database name : Fake + - View name : View2 View shape : (300, 8) + - Learning Rate : 0.7 + - Labels used : Non, Oui + - Number of cross validation folds : 5 + +Classifier configuration : + - SVM Linear with C : 2991 + - Executed on 1 core(s) + - Got configuration using randomized search with 1 iterations + + + For Accuracy score using None as sample_weights (higher is better) : + - Score on train : 0.995238095238 + - Score on test : 0.444444444444 + For F1 score using None as sample_weights, None as labels, 1 as pos_label, micro as average (higher is better) : + - Score on train : 0.995260663507 + - Score on test : 0.479166666667 + For F-beta score using None as sample_weights, None as labels, 1 as pos_label, micro as average, 1.0 as beta (higher is better) : + - Score on train : 0.995260663507 + - Score on test : 0.479166666667 + For Hamming loss using None as classes (lower is better) : + - Score on train : 0.0047619047619 + - Score on test : 0.555555555556 + For Jaccard similarity score using None as sample_weights (higher is better) : + - Score on train : 0.995238095238 + - Score on test : 0.444444444444 + For Log loss using None as sample_weights, 1e-15 as eps (lower is better) : + - Score on train : nan + - Score on test : nan + For Matthews correlation coefficient (higher is better) : + - Score on train : 0.990521113087 + - Score on test : -0.0940733030728 + For Precision score using None as sample_weights, None as labels, 1 as pos_label, binary as average (higher is better) : + - Score on train : 1.0 + - Score on test : 0.418181818182 + For Recall score using None as sample_weights, None as labels, 1 as pos_label, binary as average (higher is better) : + - Score on train : 0.990566037736 + - Score on test : 0.560975609756 + For ROS AUC score using None as sample_weights, micro as average (higher is better) : + - Score on train : 0.995283018868 + - Score on test : 0.453957192633 + For Zero one loss using None as sample_weights (lower is better) : + - Score on train : 0.0047619047619 + - Score on test : 0.555555555556 + + + Classification took 0:00:00 +2016-09-06 11:21:58,680 INFO: Done: Result Analysis +2016-09-06 11:21:58,822 DEBUG: ### Main Programm for Classification MonoView +2016-09-06 11:21:58,822 DEBUG: ### Main Programm for Classification MonoView +2016-09-06 11:21:58,822 DEBUG: ### Classification - Database:Fake Feature:View3 train_size:0.7, CrossValidation k-folds:5, cores:1, algorithm : Adaboost +2016-09-06 11:21:58,822 DEBUG: ### Classification - Database:Fake Feature:View3 train_size:0.7, CrossValidation k-folds:5, cores:1, algorithm : DecisionTree +2016-09-06 11:21:58,822 DEBUG: Start: Determine Train/Test split +2016-09-06 11:21:58,822 DEBUG: Start: Determine Train/Test split +2016-09-06 11:21:58,823 DEBUG: Info: Shape X_train:(210, 6), Length of y_train:210 +2016-09-06 11:21:58,823 DEBUG: Info: Shape X_train:(210, 6), Length of y_train:210 +2016-09-06 11:21:58,823 DEBUG: Info: Shape X_test:(90, 6), Length of y_test:90 +2016-09-06 11:21:58,823 DEBUG: Info: Shape X_test:(90, 6), Length of y_test:90 +2016-09-06 11:21:58,823 DEBUG: Done: Determine Train/Test split +2016-09-06 11:21:58,823 DEBUG: Done: Determine Train/Test split +2016-09-06 11:21:58,823 DEBUG: Start: RandomSearch best settings with 1 iterations +2016-09-06 11:21:58,823 DEBUG: Start: RandomSearch best settings with 1 iterations +2016-09-06 11:21:58,874 DEBUG: Done: RandomSearch best settings +2016-09-06 11:21:58,874 DEBUG: Start: Training +2016-09-06 11:21:58,875 DEBUG: Info: Time for Training: 0.0545539855957[s] +2016-09-06 11:21:58,876 DEBUG: Done: Training +2016-09-06 11:21:58,876 DEBUG: Start: Predicting +2016-09-06 11:21:58,880 DEBUG: Done: Predicting +2016-09-06 11:21:58,880 DEBUG: Start: Getting Results +2016-09-06 11:21:58,899 DEBUG: Done: RandomSearch best settings +2016-09-06 11:21:58,899 DEBUG: Start: Training +2016-09-06 11:21:58,905 DEBUG: Info: Time for Training: 0.0836429595947[s] +2016-09-06 11:21:58,905 DEBUG: Done: Training +2016-09-06 11:21:58,905 DEBUG: Start: Predicting +2016-09-06 11:21:58,908 DEBUG: Done: Predicting +2016-09-06 11:21:58,908 DEBUG: Start: Getting Results +2016-09-06 11:21:58,930 DEBUG: Done: Getting Results +2016-09-06 11:21:58,930 INFO: Classification on Fake database for View3 with DecisionTree + +accuracy_score on train : 1.0 +accuracy_score on test : 0.466666666667 + +Database configuration : + - Database name : Fake + - View name : View3 View shape : (300, 6) + - Learning Rate : 0.7 + - Labels used : Non, Oui + - Number of cross validation folds : 5 + +Classifier configuration : + - Decision Tree with max_depth : 15 + - Executed on 1 core(s) + - Got configuration using randomized search with 1 iterations + + + For Accuracy score using None as sample_weights (higher is better) : + - Score on train : 1.0 + - Score on test : 0.466666666667 + For F1 score using None as sample_weights, None as labels, 1 as pos_label, micro as average (higher is better) : + - Score on train : 1.0 + - Score on test : 0.428571428571 + For F-beta score using None as sample_weights, None as labels, 1 as pos_label, micro as average, 1.0 as beta (higher is better) : + - Score on train : 1.0 + - Score on test : 0.428571428571 + For Hamming loss using None as classes (lower is better) : + - Score on train : 0.0 + - Score on test : 0.533333333333 + For Jaccard similarity score using None as sample_weights (higher is better) : + - Score on train : 1.0 + - Score on test : 0.466666666667 + For Log loss using None as sample_weights, 1e-15 as eps (lower is better) : + - Score on train : nan + - Score on test : nan + For Matthews correlation coefficient (higher is better) : + - Score on train : 1.0 + - Score on test : -0.0709680565554 + For Precision score using None as sample_weights, None as labels, 1 as pos_label, binary as average (higher is better) : + - Score on train : 1.0 + - Score on test : 0.418604651163 + For Recall score using None as sample_weights, None as labels, 1 as pos_label, binary as average (higher is better) : + - Score on train : 1.0 + - Score on test : 0.439024390244 + For ROS AUC score using None as sample_weights, micro as average (higher is better) : + - Score on train : 1.0 + - Score on test : 0.464410154306 + For Zero one loss using None as sample_weights (lower is better) : + - Score on train : 0.0 + - Score on test : 0.533333333333 + + + Classification took 0:00:00 +2016-09-06 11:21:58,931 INFO: Done: Result Analysis +2016-09-06 11:21:58,947 DEBUG: Done: Getting Results +2016-09-06 11:21:58,947 INFO: Classification on Fake database for View3 with Adaboost + +accuracy_score on train : 1.0 +accuracy_score on test : 0.444444444444 + +Database configuration : + - Database name : Fake + - View name : View3 View shape : (300, 6) + - Learning Rate : 0.7 + - Labels used : Non, Oui + - Number of cross validation folds : 5 + +Classifier configuration : + - Adaboost with num_esimators : 7, base_estimators : DecisionTreeClassifier(class_weight=None, criterion='gini', max_depth=None, + max_features=None, max_leaf_nodes=None, min_samples_leaf=1, + min_samples_split=2, min_weight_fraction_leaf=0.0, + presort=False, random_state=None, splitter='best') + - Executed on 1 core(s) + - Got configuration using randomized search with 1 iterations + + + For Accuracy score using None as sample_weights (higher is better) : + - Score on train : 1.0 + - Score on test : 0.444444444444 + For F1 score using None as sample_weights, None as labels, 1 as pos_label, micro as average (higher is better) : + - Score on train : 1.0 + - Score on test : 0.404761904762 + For F-beta score using None as sample_weights, None as labels, 1 as pos_label, micro as average, 1.0 as beta (higher is better) : + - Score on train : 1.0 + - Score on test : 0.404761904762 + For Hamming loss using None as classes (lower is better) : + - Score on train : 0.0 + - Score on test : 0.555555555556 + For Jaccard similarity score using None as sample_weights (higher is better) : + - Score on train : 1.0 + - Score on test : 0.444444444444 + For Log loss using None as sample_weights, 1e-15 as eps (lower is better) : + - Score on train : nan + - Score on test : nan + For Matthews correlation coefficient (higher is better) : + - Score on train : 1.0 + - Score on test : -0.115633266975 + For Precision score using None as sample_weights, None as labels, 1 as pos_label, binary as average (higher is better) : + - Score on train : 1.0 + - Score on test : 0.395348837209 + For Recall score using None as sample_weights, None as labels, 1 as pos_label, binary as average (higher is better) : + - Score on train : 1.0 + - Score on test : 0.414634146341 + For ROS AUC score using None as sample_weights, micro as average (higher is better) : + - Score on train : 1.0 + - Score on test : 0.442010950722 + For Zero one loss using None as sample_weights (lower is better) : + - Score on train : 0.0 + - Score on test : 0.555555555556 + + + Classification took 0:00:00 +2016-09-06 11:21:58,947 INFO: Done: Result Analysis +2016-09-06 11:21:59,070 DEBUG: ### Main Programm for Classification MonoView +2016-09-06 11:21:59,070 DEBUG: ### Main Programm for Classification MonoView +2016-09-06 11:21:59,070 DEBUG: ### Classification - Database:Fake Feature:View3 train_size:0.7, CrossValidation k-folds:5, cores:1, algorithm : KNN +2016-09-06 11:21:59,070 DEBUG: ### Classification - Database:Fake Feature:View3 train_size:0.7, CrossValidation k-folds:5, cores:1, algorithm : RandomForest +2016-09-06 11:21:59,070 DEBUG: Start: Determine Train/Test split +2016-09-06 11:21:59,070 DEBUG: Start: Determine Train/Test split +2016-09-06 11:21:59,071 DEBUG: Info: Shape X_train:(210, 6), Length of y_train:210 +2016-09-06 11:21:59,071 DEBUG: Info: Shape X_train:(210, 6), Length of y_train:210 +2016-09-06 11:21:59,071 DEBUG: Info: Shape X_test:(90, 6), Length of y_test:90 +2016-09-06 11:21:59,071 DEBUG: Info: Shape X_test:(90, 6), Length of y_test:90 +2016-09-06 11:21:59,071 DEBUG: Done: Determine Train/Test split +2016-09-06 11:21:59,071 DEBUG: Done: Determine Train/Test split +2016-09-06 11:21:59,071 DEBUG: Start: RandomSearch best settings with 1 iterations +2016-09-06 11:21:59,071 DEBUG: Start: RandomSearch best settings with 1 iterations +2016-09-06 11:21:59,101 DEBUG: Done: RandomSearch best settings +2016-09-06 11:21:59,102 DEBUG: Start: Training +2016-09-06 11:21:59,102 DEBUG: Info: Time for Training: 0.0325701236725[s] +2016-09-06 11:21:59,102 DEBUG: Done: Training +2016-09-06 11:21:59,102 DEBUG: Start: Predicting +2016-09-06 11:21:59,108 DEBUG: Done: Predicting +2016-09-06 11:21:59,108 DEBUG: Start: Getting Results +2016-09-06 11:21:59,150 DEBUG: Done: Getting Results +2016-09-06 11:21:59,150 INFO: Classification on Fake database for View3 with KNN + +accuracy_score on train : 0.519047619048 +accuracy_score on test : 0.411111111111 + +Database configuration : + - Database name : Fake + - View name : View3 View shape : (300, 6) + - Learning Rate : 0.7 + - Labels used : Non, Oui + - Number of cross validation folds : 5 + +Classifier configuration : + - K nearest Neighbors with n_neighbors: 47 + - Executed on 1 core(s) + - Got configuration using randomized search with 1 iterations + + + For Accuracy score using None as sample_weights (higher is better) : + - Score on train : 0.519047619048 + - Score on test : 0.411111111111 + For F1 score using None as sample_weights, None as labels, 1 as pos_label, micro as average (higher is better) : + - Score on train : 0.562770562771 + - Score on test : 0.453608247423 + For F-beta score using None as sample_weights, None as labels, 1 as pos_label, micro as average, 1.0 as beta (higher is better) : + - Score on train : 0.562770562771 + - Score on test : 0.453608247423 + For Hamming loss using None as classes (lower is better) : + - Score on train : 0.480952380952 + - Score on test : 0.588888888889 + For Jaccard similarity score using None as sample_weights (higher is better) : + - Score on train : 0.519047619048 + - Score on test : 0.411111111111 + For Log loss using None as sample_weights, 1e-15 as eps (lower is better) : + - Score on train : nan + - Score on test : nan + For Matthews correlation coefficient (higher is better) : + - Score on train : 0.0369594857345 + - Score on test : -0.161571085301 + For Precision score using None as sample_weights, None as labels, 1 as pos_label, binary as average (higher is better) : + - Score on train : 0.52 + - Score on test : 0.392857142857 + For Recall score using None as sample_weights, None as labels, 1 as pos_label, binary as average (higher is better) : + - Score on train : 0.61320754717 + - Score on test : 0.536585365854 + For ROS AUC score using None as sample_weights, micro as average (higher is better) : + - Score on train : 0.518142235123 + - Score on test : 0.421353907417 + For Zero one loss using None as sample_weights (lower is better) : + - Score on train : 0.480952380952 + - Score on test : 0.588888888889 + + + Classification took 0:00:00 +2016-09-06 11:21:59,150 INFO: Done: Result Analysis +2016-09-06 11:21:59,228 DEBUG: Done: RandomSearch best settings +2016-09-06 11:21:59,228 DEBUG: Start: Training +2016-09-06 11:21:59,246 DEBUG: Info: Time for Training: 0.176759958267[s] +2016-09-06 11:21:59,246 DEBUG: Done: Training +2016-09-06 11:21:59,246 DEBUG: Start: Predicting +2016-09-06 11:21:59,250 DEBUG: Done: Predicting +2016-09-06 11:21:59,250 DEBUG: Start: Getting Results +2016-09-06 11:21:59,282 DEBUG: Done: Getting Results +2016-09-06 11:21:59,282 INFO: Classification on Fake database for View3 with RandomForest + +accuracy_score on train : 0.97619047619 +accuracy_score on test : 0.488888888889 + +Database configuration : + - Database name : Fake + - View name : View3 View shape : (300, 6) + - Learning Rate : 0.7 + - Labels used : Non, Oui + - Number of cross validation folds : 5 + +Classifier configuration : + - Random Forest with num_esimators : 7, max_depth : 15 + - Executed on 1 core(s) + - Got configuration using randomized search with 1 iterations + + + For Accuracy score using None as sample_weights (higher is better) : + - Score on train : 0.97619047619 + - Score on test : 0.488888888889 + For F1 score using None as sample_weights, None as labels, 1 as pos_label, micro as average (higher is better) : + - Score on train : 0.976303317536 + - Score on test : 0.410256410256 + For F-beta score using None as sample_weights, None as labels, 1 as pos_label, micro as average, 1.0 as beta (higher is better) : + - Score on train : 0.976303317536 + - Score on test : 0.410256410256 + For Hamming loss using None as classes (lower is better) : + - Score on train : 0.0238095238095 + - Score on test : 0.511111111111 + For Jaccard similarity score using None as sample_weights (higher is better) : + - Score on train : 0.97619047619 + - Score on test : 0.488888888889 + For Log loss using None as sample_weights, 1e-15 as eps (lower is better) : + - Score on train : nan + - Score on test : nan + For Matthews correlation coefficient (higher is better) : + - Score on train : 0.952424147199 + - Score on test : -0.0387937676182 + For Precision score using None as sample_weights, None as labels, 1 as pos_label, binary as average (higher is better) : + - Score on train : 0.980952380952 + - Score on test : 0.432432432432 + For Recall score using None as sample_weights, None as labels, 1 as pos_label, binary as average (higher is better) : + - Score on train : 0.971698113208 + - Score on test : 0.390243902439 + For ROS AUC score using None as sample_weights, micro as average (higher is better) : + - Score on train : 0.976233671988 + - Score on test : 0.480836236934 + For Zero one loss using None as sample_weights (lower is better) : + - Score on train : 0.0238095238095 + - Score on test : 0.511111111111 + + + Classification took 0:00:00 +2016-09-06 11:21:59,283 INFO: Done: Result Analysis +2016-09-06 11:21:59,419 DEBUG: ### Main Programm for Classification MonoView +2016-09-06 11:21:59,419 DEBUG: ### Main Programm for Classification MonoView +2016-09-06 11:21:59,419 DEBUG: ### Classification - Database:Fake Feature:View3 train_size:0.7, CrossValidation k-folds:5, cores:1, algorithm : SGD +2016-09-06 11:21:59,419 DEBUG: ### Classification - Database:Fake Feature:View3 train_size:0.7, CrossValidation k-folds:5, cores:1, algorithm : SVMLinear +2016-09-06 11:21:59,419 DEBUG: Start: Determine Train/Test split +2016-09-06 11:21:59,419 DEBUG: Start: Determine Train/Test split +2016-09-06 11:21:59,419 DEBUG: Info: Shape X_train:(210, 6), Length of y_train:210 +2016-09-06 11:21:59,419 DEBUG: Info: Shape X_train:(210, 6), Length of y_train:210 +2016-09-06 11:21:59,420 DEBUG: Info: Shape X_test:(90, 6), Length of y_test:90 +2016-09-06 11:21:59,420 DEBUG: Info: Shape X_test:(90, 6), Length of y_test:90 +2016-09-06 11:21:59,420 DEBUG: Done: Determine Train/Test split +2016-09-06 11:21:59,420 DEBUG: Done: Determine Train/Test split +2016-09-06 11:21:59,420 DEBUG: Start: RandomSearch best settings with 1 iterations +2016-09-06 11:21:59,420 DEBUG: Start: RandomSearch best settings with 1 iterations +2016-09-06 11:21:59,464 DEBUG: Done: RandomSearch best settings +2016-09-06 11:21:59,464 DEBUG: Start: Training +2016-09-06 11:21:59,465 DEBUG: Info: Time for Training: 0.0465881824493[s] +2016-09-06 11:21:59,465 DEBUG: Done: Training +2016-09-06 11:21:59,465 DEBUG: Start: Predicting +2016-09-06 11:21:59,468 DEBUG: Done: RandomSearch best settings +2016-09-06 11:21:59,468 DEBUG: Start: Training +2016-09-06 11:21:59,481 DEBUG: Done: Predicting +2016-09-06 11:21:59,481 DEBUG: Start: Getting Results +2016-09-06 11:21:59,487 DEBUG: Info: Time for Training: 0.0686330795288[s] +2016-09-06 11:21:59,487 DEBUG: Done: Training +2016-09-06 11:21:59,487 DEBUG: Start: Predicting +2016-09-06 11:21:59,490 DEBUG: Done: Predicting +2016-09-06 11:21:59,490 DEBUG: Start: Getting Results +2016-09-06 11:21:59,505 DEBUG: Done: Getting Results +2016-09-06 11:21:59,505 INFO: Classification on Fake database for View3 with SGD + +accuracy_score on train : 0.542857142857 +accuracy_score on test : 0.466666666667 + +Database configuration : + - Database name : Fake + - View name : View3 View shape : (300, 6) + - Learning Rate : 0.7 + - Labels used : Non, Oui + - Number of cross validation folds : 5 + +Classifier configuration : + - SGDClassifier with loss : modified_huber, penalty : l2 + - Executed on 1 core(s) + - Got configuration using randomized search with 1 iterations + + + For Accuracy score using None as sample_weights (higher is better) : + - Score on train : 0.542857142857 + - Score on test : 0.466666666667 + For F1 score using None as sample_weights, None as labels, 1 as pos_label, micro as average (higher is better) : + - Score on train : 0.559633027523 + - Score on test : 0.478260869565 + For F-beta score using None as sample_weights, None as labels, 1 as pos_label, micro as average, 1.0 as beta (higher is better) : + - Score on train : 0.559633027523 + - Score on test : 0.478260869565 + For Hamming loss using None as classes (lower is better) : + - Score on train : 0.457142857143 + - Score on test : 0.533333333333 + For Jaccard similarity score using None as sample_weights (higher is better) : + - Score on train : 0.542857142857 + - Score on test : 0.466666666667 + For Log loss using None as sample_weights, 1e-15 as eps (lower is better) : + - Score on train : nan + - Score on test : nan + For Matthews correlation coefficient (higher is better) : + - Score on train : 0.0852729302366 + - Score on test : -0.0555284586866 + For Precision score using None as sample_weights, None as labels, 1 as pos_label, binary as average (higher is better) : + - Score on train : 0.544642857143 + - Score on test : 0.43137254902 + For Recall score using None as sample_weights, None as labels, 1 as pos_label, binary as average (higher is better) : + - Score on train : 0.575471698113 + - Score on test : 0.536585365854 + For ROS AUC score using None as sample_weights, micro as average (higher is better) : + - Score on train : 0.542543541364 + - Score on test : 0.47237431558 + For Zero one loss using None as sample_weights (lower is better) : + - Score on train : 0.457142857143 + - Score on test : 0.533333333333 + + + Classification took 0:00:00 +2016-09-06 11:21:59,505 INFO: Done: Result Analysis +2016-09-06 11:21:59,523 DEBUG: Done: Getting Results +2016-09-06 11:21:59,523 INFO: Classification on Fake database for View3 with SVMLinear + +accuracy_score on train : 0.557142857143 +accuracy_score on test : 0.566666666667 + +Database configuration : + - Database name : Fake + - View name : View3 View shape : (300, 6) + - Learning Rate : 0.7 + - Labels used : Non, Oui + - Number of cross validation folds : 5 + +Classifier configuration : + - SVM Linear with C : 2991 + - Executed on 1 core(s) + - Got configuration using randomized search with 1 iterations + + + For Accuracy score using None as sample_weights (higher is better) : + - Score on train : 0.557142857143 + - Score on test : 0.566666666667 + For F1 score using None as sample_weights, None as labels, 1 as pos_label, micro as average (higher is better) : + - Score on train : 0.610878661088 + - Score on test : 0.571428571429 + For F-beta score using None as sample_weights, None as labels, 1 as pos_label, micro as average, 1.0 as beta (higher is better) : + - Score on train : 0.610878661088 + - Score on test : 0.571428571429 + For Hamming loss using None as classes (lower is better) : + - Score on train : 0.442857142857 + - Score on test : 0.433333333333 + For Jaccard similarity score using None as sample_weights (higher is better) : + - Score on train : 0.557142857143 + - Score on test : 0.566666666667 + For Log loss using None as sample_weights, 1e-15 as eps (lower is better) : + - Score on train : nan + - Score on test : nan + For Matthews correlation coefficient (higher is better) : + - Score on train : 0.11594977827 + - Score on test : 0.144674846981 + For Precision score using None as sample_weights, None as labels, 1 as pos_label, binary as average (higher is better) : + - Score on train : 0.548872180451 + - Score on test : 0.52 + For Recall score using None as sample_weights, None as labels, 1 as pos_label, binary as average (higher is better) : + - Score on train : 0.688679245283 + - Score on test : 0.634146341463 + For ROS AUC score using None as sample_weights, micro as average (higher is better) : + - Score on train : 0.55587808418 + - Score on test : 0.572175211548 + For Zero one loss using None as sample_weights (lower is better) : + - Score on train : 0.442857142857 + - Score on test : 0.433333333333 + + + Classification took 0:00:00 +2016-09-06 11:21:59,523 INFO: Done: Result Analysis +2016-09-06 11:21:59,818 INFO: ### Main Programm for Multiview Classification +2016-09-06 11:21:59,818 INFO: ### Classification - Database : Fake ; Views : Methyl, MiRNA_, RNASeq, Clinic ; Algorithm : Mumbo ; Cores : 1 +2016-09-06 11:21:59,819 INFO: Info: Shape of View0 :(300, 13) +2016-09-06 11:21:59,819 INFO: Info: Shape of View1 :(300, 18) +2016-09-06 11:21:59,820 INFO: Info: Shape of View2 :(300, 8) +2016-09-06 11:21:59,820 INFO: Info: Shape of View3 :(300, 6) +2016-09-06 11:21:59,820 INFO: Done: Read Database Files +2016-09-06 11:21:59,820 INFO: Start: Determine validation split for ratio 0.7 +2016-09-06 11:21:59,823 INFO: ### Main Programm for Multiview Classification +2016-09-06 11:21:59,823 INFO: ### Classification - Database : Fake ; Views : Methyl, MiRNA_, RNASeq, Clinic ; Algorithm : Fusion ; Cores : 1 +2016-09-06 11:21:59,823 INFO: Info: Shape of View0 :(300, 13) +2016-09-06 11:21:59,824 INFO: Info: Shape of View1 :(300, 18) +2016-09-06 11:21:59,824 INFO: Done: Determine validation split +2016-09-06 11:21:59,824 INFO: Start: Determine 5 folds +2016-09-06 11:21:59,824 INFO: Info: Shape of View2 :(300, 8) +2016-09-06 11:21:59,825 INFO: Info: Shape of View3 :(300, 6) +2016-09-06 11:21:59,825 INFO: Done: Read Database Files +2016-09-06 11:21:59,825 INFO: Start: Determine validation split for ratio 0.7 +2016-09-06 11:21:59,829 INFO: Done: Determine validation split +2016-09-06 11:21:59,829 INFO: Start: Determine 5 folds +2016-09-06 11:21:59,830 INFO: Info: Length of Learning Sets: 170 +2016-09-06 11:21:59,831 INFO: Info: Length of Testing Sets: 41 +2016-09-06 11:21:59,831 INFO: Info: Length of Validation Set: 89 +2016-09-06 11:21:59,831 INFO: Done: Determine folds +2016-09-06 11:21:59,831 INFO: Start: Learning with Mumbo and 5 folds +2016-09-06 11:21:59,831 INFO: Start: Randomsearching best settings for monoview classifiers +2016-09-06 11:21:59,831 DEBUG: Start: Gridsearch for DecisionTree on View0 +2016-09-06 11:21:59,834 INFO: Info: Length of Learning Sets: 170 +2016-09-06 11:21:59,834 INFO: Info: Length of Testing Sets: 41 +2016-09-06 11:21:59,834 INFO: Info: Length of Validation Set: 89 +2016-09-06 11:21:59,834 INFO: Done: Determine folds +2016-09-06 11:21:59,834 INFO: Start: Learning with Fusion and 5 folds +2016-09-06 11:21:59,834 INFO: Start: Randomsearching best settings for monoview classifiers +2016-09-06 11:21:59,834 DEBUG: Start: Random search for Adaboost with 30 iterations diff --git a/Code/MonoMutliViewClassifiers/Results/20160906-112155Results-Adaboost-Non-Oui-learnRate0.7-Fake.txt b/Code/MonoMutliViewClassifiers/Results/20160906-112155Results-Adaboost-Non-Oui-learnRate0.7-Fake.txt new file mode 100644 index 0000000000000000000000000000000000000000..7310aa13413c0fe51d30595a0611fbf8868e4fab --- /dev/null +++ b/Code/MonoMutliViewClassifiers/Results/20160906-112155Results-Adaboost-Non-Oui-learnRate0.7-Fake.txt @@ -0,0 +1,57 @@ +Classification on Fake database for View0 with Adaboost + +accuracy_score on train : 1.0 +accuracy_score on test : 0.5 + +Database configuration : + - Database name : Fake + - View name : View0 View shape : (300, 13) + - Learning Rate : 0.7 + - Labels used : Non, Oui + - Number of cross validation folds : 5 + +Classifier configuration : + - Adaboost with num_esimators : 7, base_estimators : DecisionTreeClassifier(class_weight=None, criterion='gini', max_depth=None, + max_features=None, max_leaf_nodes=None, min_samples_leaf=1, + min_samples_split=2, min_weight_fraction_leaf=0.0, + presort=False, random_state=None, splitter='best') + - Executed on 1 core(s) + - Got configuration using randomized search with 1 iterations + + + For Accuracy score using None as sample_weights (higher is better) : + - Score on train : 1.0 + - Score on test : 0.5 + For F1 score using None as sample_weights, None as labels, 1 as pos_label, micro as average (higher is better) : + - Score on train : 1.0 + - Score on test : 0.494382022472 + For F-beta score using None as sample_weights, None as labels, 1 as pos_label, micro as average, 1.0 as beta (higher is better) : + - Score on train : 1.0 + - Score on test : 0.494382022472 + For Hamming loss using None as classes (lower is better) : + - Score on train : 0.0 + - Score on test : 0.5 + For Jaccard similarity score using None as sample_weights (higher is better) : + - Score on train : 1.0 + - Score on test : 0.5 + For Log loss using None as sample_weights, 1e-15 as eps (lower is better) : + - Score on train : nan + - Score on test : nan + For Matthews correlation coefficient (higher is better) : + - Score on train : 1.0 + - Score on test : 0.00596274193664 + For Precision score using None as sample_weights, None as labels, 1 as pos_label, binary as average (higher is better) : + - Score on train : 1.0 + - Score on test : 0.458333333333 + For Recall score using None as sample_weights, None as labels, 1 as pos_label, binary as average (higher is better) : + - Score on train : 1.0 + - Score on test : 0.536585365854 + For ROS AUC score using None as sample_weights, micro as average (higher is better) : + - Score on train : 1.0 + - Score on test : 0.502986560478 + For Zero one loss using None as sample_weights (lower is better) : + - Score on train : 0.0 + - Score on test : 0.5 + + + Classification took 0:00:00 \ No newline at end of file diff --git a/Code/MonoMutliViewClassifiers/Results/20160906-112155Results-DecisionTree-Non-Oui-learnRate0.7-Fake.txt b/Code/MonoMutliViewClassifiers/Results/20160906-112155Results-DecisionTree-Non-Oui-learnRate0.7-Fake.txt new file mode 100644 index 0000000000000000000000000000000000000000..60733f7f4372d2fbe0d01db2bf0ef9f663fe2087 --- /dev/null +++ b/Code/MonoMutliViewClassifiers/Results/20160906-112155Results-DecisionTree-Non-Oui-learnRate0.7-Fake.txt @@ -0,0 +1,54 @@ +Classification on Fake database for View0 with DecisionTree + +accuracy_score on train : 1.0 +accuracy_score on test : 0.433333333333 + +Database configuration : + - Database name : Fake + - View name : View0 View shape : (300, 13) + - Learning Rate : 0.7 + - Labels used : Non, Oui + - Number of cross validation folds : 5 + +Classifier configuration : + - Decision Tree with max_depth : 15 + - Executed on 1 core(s) + - Got configuration using randomized search with 1 iterations + + + For Accuracy score using None as sample_weights (higher is better) : + - Score on train : 1.0 + - Score on test : 0.433333333333 + For F1 score using None as sample_weights, None as labels, 1 as pos_label, micro as average (higher is better) : + - Score on train : 1.0 + - Score on test : 0.413793103448 + For F-beta score using None as sample_weights, None as labels, 1 as pos_label, micro as average, 1.0 as beta (higher is better) : + - Score on train : 1.0 + - Score on test : 0.413793103448 + For Hamming loss using None as classes (lower is better) : + - Score on train : 0.0 + - Score on test : 0.566666666667 + For Jaccard similarity score using None as sample_weights (higher is better) : + - Score on train : 1.0 + - Score on test : 0.433333333333 + For Log loss using None as sample_weights, 1e-15 as eps (lower is better) : + - Score on train : nan + - Score on test : nan + For Matthews correlation coefficient (higher is better) : + - Score on train : 1.0 + - Score on test : -0.131912640639 + For Precision score using None as sample_weights, None as labels, 1 as pos_label, binary as average (higher is better) : + - Score on train : 1.0 + - Score on test : 0.391304347826 + For Recall score using None as sample_weights, None as labels, 1 as pos_label, binary as average (higher is better) : + - Score on train : 1.0 + - Score on test : 0.439024390244 + For ROS AUC score using None as sample_weights, micro as average (higher is better) : + - Score on train : 1.0 + - Score on test : 0.433797909408 + For Zero one loss using None as sample_weights (lower is better) : + - Score on train : 0.0 + - Score on test : 0.566666666667 + + + Classification took 0:00:00 \ No newline at end of file diff --git a/Code/MonoMutliViewClassifiers/Results/20160906-112155Results-KNN-Non-Oui-learnRate0.7-Fake.txt b/Code/MonoMutliViewClassifiers/Results/20160906-112155Results-KNN-Non-Oui-learnRate0.7-Fake.txt new file mode 100644 index 0000000000000000000000000000000000000000..98c9ed2a5aec9b146da2753a524452b44e1a9d01 --- /dev/null +++ b/Code/MonoMutliViewClassifiers/Results/20160906-112155Results-KNN-Non-Oui-learnRate0.7-Fake.txt @@ -0,0 +1,54 @@ +Classification on Fake database for View0 with KNN + +accuracy_score on train : 0.557142857143 +accuracy_score on test : 0.422222222222 + +Database configuration : + - Database name : Fake + - View name : View0 View shape : (300, 13) + - Learning Rate : 0.7 + - Labels used : Non, Oui + - Number of cross validation folds : 5 + +Classifier configuration : + - K nearest Neighbors with n_neighbors: 47 + - Executed on 1 core(s) + - Got configuration using randomized search with 1 iterations + + + For Accuracy score using None as sample_weights (higher is better) : + - Score on train : 0.557142857143 + - Score on test : 0.422222222222 + For F1 score using None as sample_weights, None as labels, 1 as pos_label, micro as average (higher is better) : + - Score on train : 0.550724637681 + - Score on test : 0.315789473684 + For F-beta score using None as sample_weights, None as labels, 1 as pos_label, micro as average, 1.0 as beta (higher is better) : + - Score on train : 0.550724637681 + - Score on test : 0.315789473684 + For Hamming loss using None as classes (lower is better) : + - Score on train : 0.442857142857 + - Score on test : 0.577777777778 + For Jaccard similarity score using None as sample_weights (higher is better) : + - Score on train : 0.557142857143 + - Score on test : 0.422222222222 + For Log loss using None as sample_weights, 1e-15 as eps (lower is better) : + - Score on train : nan + - Score on test : nan + For Matthews correlation coefficient (higher is better) : + - Score on train : 0.11473701202 + - Score on test : -0.180519041032 + For Precision score using None as sample_weights, None as labels, 1 as pos_label, binary as average (higher is better) : + - Score on train : 0.564356435644 + - Score on test : 0.342857142857 + For Recall score using None as sample_weights, None as labels, 1 as pos_label, binary as average (higher is better) : + - Score on train : 0.537735849057 + - Score on test : 0.292682926829 + For ROS AUC score using None as sample_weights, micro as average (higher is better) : + - Score on train : 0.55732946299 + - Score on test : 0.411647585864 + For Zero one loss using None as sample_weights (lower is better) : + - Score on train : 0.442857142857 + - Score on test : 0.577777777778 + + + Classification took 0:00:00 \ No newline at end of file diff --git a/Code/MonoMutliViewClassifiers/Results/20160906-112155Results-RandomForest-Non-Oui-learnRate0.7-Fake.txt b/Code/MonoMutliViewClassifiers/Results/20160906-112155Results-RandomForest-Non-Oui-learnRate0.7-Fake.txt new file mode 100644 index 0000000000000000000000000000000000000000..11143f3f7b7d16b5e170c5ab4fd7f72ea6e130b2 --- /dev/null +++ b/Code/MonoMutliViewClassifiers/Results/20160906-112155Results-RandomForest-Non-Oui-learnRate0.7-Fake.txt @@ -0,0 +1,54 @@ +Classification on Fake database for View0 with RandomForest + +accuracy_score on train : 0.957142857143 +accuracy_score on test : 0.4 + +Database configuration : + - Database name : Fake + - View name : View0 View shape : (300, 13) + - Learning Rate : 0.7 + - Labels used : Non, Oui + - Number of cross validation folds : 5 + +Classifier configuration : + - Random Forest with num_esimators : 7, max_depth : 15 + - Executed on 1 core(s) + - Got configuration using randomized search with 1 iterations + + + For Accuracy score using None as sample_weights (higher is better) : + - Score on train : 0.957142857143 + - Score on test : 0.4 + For F1 score using None as sample_weights, None as labels, 1 as pos_label, micro as average (higher is better) : + - Score on train : 0.956937799043 + - Score on test : 0.325 + For F-beta score using None as sample_weights, None as labels, 1 as pos_label, micro as average, 1.0 as beta (higher is better) : + - Score on train : 0.956937799043 + - Score on test : 0.325 + For Hamming loss using None as classes (lower is better) : + - Score on train : 0.0428571428571 + - Score on test : 0.6 + For Jaccard similarity score using None as sample_weights (higher is better) : + - Score on train : 0.957142857143 + - Score on test : 0.4 + For Log loss using None as sample_weights, 1e-15 as eps (lower is better) : + - Score on train : nan + - Score on test : nan + For Matthews correlation coefficient (higher is better) : + - Score on train : 0.914674537841 + - Score on test : -0.214609988978 + For Precision score using None as sample_weights, None as labels, 1 as pos_label, binary as average (higher is better) : + - Score on train : 0.970873786408 + - Score on test : 0.333333333333 + For Recall score using None as sample_weights, None as labels, 1 as pos_label, binary as average (higher is better) : + - Score on train : 0.943396226415 + - Score on test : 0.317073170732 + For ROS AUC score using None as sample_weights, micro as average (higher is better) : + - Score on train : 0.957275036284 + - Score on test : 0.393230462917 + For Zero one loss using None as sample_weights (lower is better) : + - Score on train : 0.0428571428571 + - Score on test : 0.6 + + + Classification took 0:00:00 \ No newline at end of file diff --git a/Code/MonoMutliViewClassifiers/Results/20160906-112156Results-Adaboost-Non-Oui-learnRate0.7-Fake.txt b/Code/MonoMutliViewClassifiers/Results/20160906-112156Results-Adaboost-Non-Oui-learnRate0.7-Fake.txt new file mode 100644 index 0000000000000000000000000000000000000000..22da83be580cf08f7fe511bc720910ca2202c3f6 --- /dev/null +++ b/Code/MonoMutliViewClassifiers/Results/20160906-112156Results-Adaboost-Non-Oui-learnRate0.7-Fake.txt @@ -0,0 +1,57 @@ +Classification on Fake database for View1 with Adaboost + +accuracy_score on train : 1.0 +accuracy_score on test : 0.444444444444 + +Database configuration : + - Database name : Fake + - View name : View1 View shape : (300, 18) + - Learning Rate : 0.7 + - Labels used : Non, Oui + - Number of cross validation folds : 5 + +Classifier configuration : + - Adaboost with num_esimators : 7, base_estimators : DecisionTreeClassifier(class_weight=None, criterion='gini', max_depth=None, + max_features=None, max_leaf_nodes=None, min_samples_leaf=1, + min_samples_split=2, min_weight_fraction_leaf=0.0, + presort=False, random_state=None, splitter='best') + - Executed on 1 core(s) + - Got configuration using randomized search with 1 iterations + + + For Accuracy score using None as sample_weights (higher is better) : + - Score on train : 1.0 + - Score on test : 0.444444444444 + For F1 score using None as sample_weights, None as labels, 1 as pos_label, micro as average (higher is better) : + - Score on train : 1.0 + - Score on test : 0.404761904762 + For F-beta score using None as sample_weights, None as labels, 1 as pos_label, micro as average, 1.0 as beta (higher is better) : + - Score on train : 1.0 + - Score on test : 0.404761904762 + For Hamming loss using None as classes (lower is better) : + - Score on train : 0.0 + - Score on test : 0.555555555556 + For Jaccard similarity score using None as sample_weights (higher is better) : + - Score on train : 1.0 + - Score on test : 0.444444444444 + For Log loss using None as sample_weights, 1e-15 as eps (lower is better) : + - Score on train : nan + - Score on test : nan + For Matthews correlation coefficient (higher is better) : + - Score on train : 1.0 + - Score on test : -0.115633266975 + For Precision score using None as sample_weights, None as labels, 1 as pos_label, binary as average (higher is better) : + - Score on train : 1.0 + - Score on test : 0.395348837209 + For Recall score using None as sample_weights, None as labels, 1 as pos_label, binary as average (higher is better) : + - Score on train : 1.0 + - Score on test : 0.414634146341 + For ROS AUC score using None as sample_weights, micro as average (higher is better) : + - Score on train : 1.0 + - Score on test : 0.442010950722 + For Zero one loss using None as sample_weights (lower is better) : + - Score on train : 0.0 + - Score on test : 0.555555555556 + + + Classification took 0:00:00 \ No newline at end of file diff --git a/Code/MonoMutliViewClassifiers/Results/20160906-112156Results-DecisionTree-Non-Oui-learnRate0.7-Fake.txt b/Code/MonoMutliViewClassifiers/Results/20160906-112156Results-DecisionTree-Non-Oui-learnRate0.7-Fake.txt new file mode 100644 index 0000000000000000000000000000000000000000..97bcfb03eede2f30c7c09839dcbc52d8578c8b88 --- /dev/null +++ b/Code/MonoMutliViewClassifiers/Results/20160906-112156Results-DecisionTree-Non-Oui-learnRate0.7-Fake.txt @@ -0,0 +1,54 @@ +Classification on Fake database for View1 with DecisionTree + +accuracy_score on train : 1.0 +accuracy_score on test : 0.444444444444 + +Database configuration : + - Database name : Fake + - View name : View1 View shape : (300, 18) + - Learning Rate : 0.7 + - Labels used : Non, Oui + - Number of cross validation folds : 5 + +Classifier configuration : + - Decision Tree with max_depth : 15 + - Executed on 1 core(s) + - Got configuration using randomized search with 1 iterations + + + For Accuracy score using None as sample_weights (higher is better) : + - Score on train : 1.0 + - Score on test : 0.444444444444 + For F1 score using None as sample_weights, None as labels, 1 as pos_label, micro as average (higher is better) : + - Score on train : 1.0 + - Score on test : 0.418604651163 + For F-beta score using None as sample_weights, None as labels, 1 as pos_label, micro as average, 1.0 as beta (higher is better) : + - Score on train : 1.0 + - Score on test : 0.418604651163 + For Hamming loss using None as classes (lower is better) : + - Score on train : 0.0 + - Score on test : 0.555555555556 + For Jaccard similarity score using None as sample_weights (higher is better) : + - Score on train : 1.0 + - Score on test : 0.444444444444 + For Log loss using None as sample_weights, 1e-15 as eps (lower is better) : + - Score on train : nan + - Score on test : nan + For Matthews correlation coefficient (higher is better) : + - Score on train : 1.0 + - Score on test : -0.111552687063 + For Precision score using None as sample_weights, None as labels, 1 as pos_label, binary as average (higher is better) : + - Score on train : 1.0 + - Score on test : 0.4 + For Recall score using None as sample_weights, None as labels, 1 as pos_label, binary as average (higher is better) : + - Score on train : 1.0 + - Score on test : 0.439024390244 + For ROS AUC score using None as sample_weights, micro as average (higher is better) : + - Score on train : 1.0 + - Score on test : 0.44400199104 + For Zero one loss using None as sample_weights (lower is better) : + - Score on train : 0.0 + - Score on test : 0.555555555556 + + + Classification took 0:00:00 \ No newline at end of file diff --git a/Code/MonoMutliViewClassifiers/Results/20160906-112156Results-KNN-Non-Oui-learnRate0.7-Fake.txt b/Code/MonoMutliViewClassifiers/Results/20160906-112156Results-KNN-Non-Oui-learnRate0.7-Fake.txt new file mode 100644 index 0000000000000000000000000000000000000000..db1e7c90d39f84e92df0500245ea9b9e26d1dedb --- /dev/null +++ b/Code/MonoMutliViewClassifiers/Results/20160906-112156Results-KNN-Non-Oui-learnRate0.7-Fake.txt @@ -0,0 +1,54 @@ +Classification on Fake database for View1 with KNN + +accuracy_score on train : 0.547619047619 +accuracy_score on test : 0.455555555556 + +Database configuration : + - Database name : Fake + - View name : View1 View shape : (300, 18) + - Learning Rate : 0.7 + - Labels used : Non, Oui + - Number of cross validation folds : 5 + +Classifier configuration : + - K nearest Neighbors with n_neighbors: 47 + - Executed on 1 core(s) + - Got configuration using randomized search with 1 iterations + + + For Accuracy score using None as sample_weights (higher is better) : + - Score on train : 0.547619047619 + - Score on test : 0.455555555556 + For F1 score using None as sample_weights, None as labels, 1 as pos_label, micro as average (higher is better) : + - Score on train : 0.633204633205 + - Score on test : 0.558558558559 + For F-beta score using None as sample_weights, None as labels, 1 as pos_label, micro as average, 1.0 as beta (higher is better) : + - Score on train : 0.633204633205 + - Score on test : 0.558558558559 + For Hamming loss using None as classes (lower is better) : + - Score on train : 0.452380952381 + - Score on test : 0.544444444444 + For Jaccard similarity score using None as sample_weights (higher is better) : + - Score on train : 0.547619047619 + - Score on test : 0.455555555556 + For Log loss using None as sample_weights, 1e-15 as eps (lower is better) : + - Score on train : nan + - Score on test : nan + For Matthews correlation coefficient (higher is better) : + - Score on train : 0.102191547553 + - Score on test : -0.0477019354931 + For Precision score using None as sample_weights, None as labels, 1 as pos_label, binary as average (higher is better) : + - Score on train : 0.535947712418 + - Score on test : 0.442857142857 + For Recall score using None as sample_weights, None as labels, 1 as pos_label, binary as average (higher is better) : + - Score on train : 0.77358490566 + - Score on test : 0.756097560976 + For ROS AUC score using None as sample_weights, micro as average (higher is better) : + - Score on train : 0.545446298984 + - Score on test : 0.480089596814 + For Zero one loss using None as sample_weights (lower is better) : + - Score on train : 0.452380952381 + - Score on test : 0.544444444444 + + + Classification took 0:00:00 \ No newline at end of file diff --git a/Code/MonoMutliViewClassifiers/Results/20160906-112156Results-SGD-Non-Oui-learnRate0.7-Fake.txt b/Code/MonoMutliViewClassifiers/Results/20160906-112156Results-SGD-Non-Oui-learnRate0.7-Fake.txt new file mode 100644 index 0000000000000000000000000000000000000000..52272501df553f770e104ca4715e1eb70fe7a93f --- /dev/null +++ b/Code/MonoMutliViewClassifiers/Results/20160906-112156Results-SGD-Non-Oui-learnRate0.7-Fake.txt @@ -0,0 +1,54 @@ +Classification on Fake database for View0 with SGD + +accuracy_score on train : 0.604761904762 +accuracy_score on test : 0.433333333333 + +Database configuration : + - Database name : Fake + - View name : View0 View shape : (300, 13) + - Learning Rate : 0.7 + - Labels used : Non, Oui + - Number of cross validation folds : 5 + +Classifier configuration : + - SGDClassifier with loss : modified_huber, penalty : l2 + - Executed on 1 core(s) + - Got configuration using randomized search with 1 iterations + + + For Accuracy score using None as sample_weights (higher is better) : + - Score on train : 0.604761904762 + - Score on test : 0.433333333333 + For F1 score using None as sample_weights, None as labels, 1 as pos_label, micro as average (higher is better) : + - Score on train : 0.62443438914 + - Score on test : 0.43956043956 + For F-beta score using None as sample_weights, None as labels, 1 as pos_label, micro as average, 1.0 as beta (higher is better) : + - Score on train : 0.62443438914 + - Score on test : 0.43956043956 + For Hamming loss using None as classes (lower is better) : + - Score on train : 0.395238095238 + - Score on test : 0.566666666667 + For Jaccard similarity score using None as sample_weights (higher is better) : + - Score on train : 0.604761904762 + - Score on test : 0.433333333333 + For Log loss using None as sample_weights, 1e-15 as eps (lower is better) : + - Score on train : nan + - Score on test : nan + For Matthews correlation coefficient (higher is better) : + - Score on train : 0.209578877963 + - Score on test : -0.124719695673 + For Precision score using None as sample_weights, None as labels, 1 as pos_label, binary as average (higher is better) : + - Score on train : 0.6 + - Score on test : 0.4 + For Recall score using None as sample_weights, None as labels, 1 as pos_label, binary as average (higher is better) : + - Score on train : 0.650943396226 + - Score on test : 0.487804878049 + For ROS AUC score using None as sample_weights, micro as average (higher is better) : + - Score on train : 0.604317851959 + - Score on test : 0.437779990045 + For Zero one loss using None as sample_weights (lower is better) : + - Score on train : 0.395238095238 + - Score on test : 0.566666666667 + + + Classification took 0:00:00 \ No newline at end of file diff --git a/Code/MonoMutliViewClassifiers/Results/20160906-112156Results-SVMLinear-Non-Oui-learnRate0.7-Fake.txt b/Code/MonoMutliViewClassifiers/Results/20160906-112156Results-SVMLinear-Non-Oui-learnRate0.7-Fake.txt new file mode 100644 index 0000000000000000000000000000000000000000..023c955b3646997254d66b6f74a22072139d0cdc --- /dev/null +++ b/Code/MonoMutliViewClassifiers/Results/20160906-112156Results-SVMLinear-Non-Oui-learnRate0.7-Fake.txt @@ -0,0 +1,54 @@ +Classification on Fake database for View0 with SVMLinear + +accuracy_score on train : 0.495238095238 +accuracy_score on test : 0.444444444444 + +Database configuration : + - Database name : Fake + - View name : View0 View shape : (300, 13) + - Learning Rate : 0.7 + - Labels used : Non, Oui + - Number of cross validation folds : 5 + +Classifier configuration : + - SVM Linear with C : 2991 + - Executed on 1 core(s) + - Got configuration using randomized search with 1 iterations + + + For Accuracy score using None as sample_weights (higher is better) : + - Score on train : 0.495238095238 + - Score on test : 0.444444444444 + For F1 score using None as sample_weights, None as labels, 1 as pos_label, micro as average (higher is better) : + - Score on train : 0.504672897196 + - Score on test : 0.404761904762 + For F-beta score using None as sample_weights, None as labels, 1 as pos_label, micro as average, 1.0 as beta (higher is better) : + - Score on train : 0.504672897196 + - Score on test : 0.404761904762 + For Hamming loss using None as classes (lower is better) : + - Score on train : 0.504761904762 + - Score on test : 0.555555555556 + For Jaccard similarity score using None as sample_weights (higher is better) : + - Score on train : 0.495238095238 + - Score on test : 0.444444444444 + For Log loss using None as sample_weights, 1e-15 as eps (lower is better) : + - Score on train : nan + - Score on test : nan + For Matthews correlation coefficient (higher is better) : + - Score on train : -0.00980036362201 + - Score on test : -0.115633266975 + For Precision score using None as sample_weights, None as labels, 1 as pos_label, binary as average (higher is better) : + - Score on train : 0.5 + - Score on test : 0.395348837209 + For Recall score using None as sample_weights, None as labels, 1 as pos_label, binary as average (higher is better) : + - Score on train : 0.509433962264 + - Score on test : 0.414634146341 + For ROS AUC score using None as sample_weights, micro as average (higher is better) : + - Score on train : 0.495101596517 + - Score on test : 0.442010950722 + For Zero one loss using None as sample_weights (lower is better) : + - Score on train : 0.504761904762 + - Score on test : 0.555555555556 + + + Classification took 0:00:00 \ No newline at end of file diff --git a/Code/MonoMutliViewClassifiers/Results/20160906-112156Results-SVMPoly-Non-Oui-learnRate0.7-Fake.txt b/Code/MonoMutliViewClassifiers/Results/20160906-112156Results-SVMPoly-Non-Oui-learnRate0.7-Fake.txt new file mode 100644 index 0000000000000000000000000000000000000000..4a817800d802fe8d5147dc163ddbc8afb6cdbb72 --- /dev/null +++ b/Code/MonoMutliViewClassifiers/Results/20160906-112156Results-SVMPoly-Non-Oui-learnRate0.7-Fake.txt @@ -0,0 +1,54 @@ +Classification on Fake database for View0 with SVMPoly + +accuracy_score on train : 1.0 +accuracy_score on test : 0.5 + +Database configuration : + - Database name : Fake + - View name : View0 View shape : (300, 13) + - Learning Rate : 0.7 + - Labels used : Non, Oui + - Number of cross validation folds : 5 + +Classifier configuration : + - SVM Linear with C : 2991 + - Executed on 1 core(s) + - Got configuration using randomized search with 1 iterations + + + For Accuracy score using None as sample_weights (higher is better) : + - Score on train : 1.0 + - Score on test : 0.5 + For F1 score using None as sample_weights, None as labels, 1 as pos_label, micro as average (higher is better) : + - Score on train : 1.0 + - Score on test : 0.150943396226 + For F-beta score using None as sample_weights, None as labels, 1 as pos_label, micro as average, 1.0 as beta (higher is better) : + - Score on train : 1.0 + - Score on test : 0.150943396226 + For Hamming loss using None as classes (lower is better) : + - Score on train : 0.0 + - Score on test : 0.5 + For Jaccard similarity score using None as sample_weights (higher is better) : + - Score on train : 1.0 + - Score on test : 0.5 + For Log loss using None as sample_weights, 1e-15 as eps (lower is better) : + - Score on train : nan + - Score on test : nan + For Matthews correlation coefficient (higher is better) : + - Score on train : 1.0 + - Score on test : -0.096260040145 + For Precision score using None as sample_weights, None as labels, 1 as pos_label, binary as average (higher is better) : + - Score on train : 1.0 + - Score on test : 0.333333333333 + For Recall score using None as sample_weights, None as labels, 1 as pos_label, binary as average (higher is better) : + - Score on train : 1.0 + - Score on test : 0.0975609756098 + For ROS AUC score using None as sample_weights, micro as average (higher is better) : + - Score on train : 1.0 + - Score on test : 0.467147834744 + For Zero one loss using None as sample_weights (lower is better) : + - Score on train : 0.0 + - Score on test : 0.5 + + + Classification took 0:00:00 \ No newline at end of file diff --git a/Code/MonoMutliViewClassifiers/Results/20160906-112156Results-SVMRBF-Non-Oui-learnRate0.7-Fake.txt b/Code/MonoMutliViewClassifiers/Results/20160906-112156Results-SVMRBF-Non-Oui-learnRate0.7-Fake.txt new file mode 100644 index 0000000000000000000000000000000000000000..9c326ee0a2e05dea68d6e24bbb431c6b26d3dc19 --- /dev/null +++ b/Code/MonoMutliViewClassifiers/Results/20160906-112156Results-SVMRBF-Non-Oui-learnRate0.7-Fake.txt @@ -0,0 +1,54 @@ +Classification on Fake database for View0 with SVMRBF + +accuracy_score on train : 1.0 +accuracy_score on test : 0.411111111111 + +Database configuration : + - Database name : Fake + - View name : View0 View shape : (300, 13) + - Learning Rate : 0.7 + - Labels used : Non, Oui + - Number of cross validation folds : 5 + +Classifier configuration : + - SVM Linear with C : 2991 + - Executed on 1 core(s) + - Got configuration using randomized search with 1 iterations + + + For Accuracy score using None as sample_weights (higher is better) : + - Score on train : 1.0 + - Score on test : 0.411111111111 + For F1 score using None as sample_weights, None as labels, 1 as pos_label, micro as average (higher is better) : + - Score on train : 1.0 + - Score on test : 0.345679012346 + For F-beta score using None as sample_weights, None as labels, 1 as pos_label, micro as average, 1.0 as beta (higher is better) : + - Score on train : 1.0 + - Score on test : 0.345679012346 + For Hamming loss using None as classes (lower is better) : + - Score on train : 0.0 + - Score on test : 0.588888888889 + For Jaccard similarity score using None as sample_weights (higher is better) : + - Score on train : 1.0 + - Score on test : 0.411111111111 + For Log loss using None as sample_weights, 1e-15 as eps (lower is better) : + - Score on train : nan + - Score on test : nan + For Matthews correlation coefficient (higher is better) : + - Score on train : 1.0 + - Score on test : -0.189573937423 + For Precision score using None as sample_weights, None as labels, 1 as pos_label, binary as average (higher is better) : + - Score on train : 1.0 + - Score on test : 0.35 + For Recall score using None as sample_weights, None as labels, 1 as pos_label, binary as average (higher is better) : + - Score on train : 1.0 + - Score on test : 0.341463414634 + For ROS AUC score using None as sample_weights, micro as average (higher is better) : + - Score on train : 1.0 + - Score on test : 0.405425584868 + For Zero one loss using None as sample_weights (lower is better) : + - Score on train : 0.0 + - Score on test : 0.588888888889 + + + Classification took 0:00:00 \ No newline at end of file diff --git a/Code/MonoMutliViewClassifiers/Results/20160906-112157Results-Adaboost-Non-Oui-learnRate0.7-Fake.txt b/Code/MonoMutliViewClassifiers/Results/20160906-112157Results-Adaboost-Non-Oui-learnRate0.7-Fake.txt new file mode 100644 index 0000000000000000000000000000000000000000..bc43f864d2ea828df712fbae60f20043f68cfb9c --- /dev/null +++ b/Code/MonoMutliViewClassifiers/Results/20160906-112157Results-Adaboost-Non-Oui-learnRate0.7-Fake.txt @@ -0,0 +1,57 @@ +Classification on Fake database for View2 with Adaboost + +accuracy_score on train : 1.0 +accuracy_score on test : 0.555555555556 + +Database configuration : + - Database name : Fake + - View name : View2 View shape : (300, 8) + - Learning Rate : 0.7 + - Labels used : Non, Oui + - Number of cross validation folds : 5 + +Classifier configuration : + - Adaboost with num_esimators : 7, base_estimators : DecisionTreeClassifier(class_weight=None, criterion='gini', max_depth=None, + max_features=None, max_leaf_nodes=None, min_samples_leaf=1, + min_samples_split=2, min_weight_fraction_leaf=0.0, + presort=False, random_state=None, splitter='best') + - Executed on 1 core(s) + - Got configuration using randomized search with 1 iterations + + + For Accuracy score using None as sample_weights (higher is better) : + - Score on train : 1.0 + - Score on test : 0.555555555556 + For F1 score using None as sample_weights, None as labels, 1 as pos_label, micro as average (higher is better) : + - Score on train : 1.0 + - Score on test : 0.512195121951 + For F-beta score using None as sample_weights, None as labels, 1 as pos_label, micro as average, 1.0 as beta (higher is better) : + - Score on train : 1.0 + - Score on test : 0.512195121951 + For Hamming loss using None as classes (lower is better) : + - Score on train : 0.0 + - Score on test : 0.444444444444 + For Jaccard similarity score using None as sample_weights (higher is better) : + - Score on train : 1.0 + - Score on test : 0.555555555556 + For Log loss using None as sample_weights, 1e-15 as eps (lower is better) : + - Score on train : nan + - Score on test : nan + For Matthews correlation coefficient (higher is better) : + - Score on train : 1.0 + - Score on test : 0.104031856645 + For Precision score using None as sample_weights, None as labels, 1 as pos_label, binary as average (higher is better) : + - Score on train : 1.0 + - Score on test : 0.512195121951 + For Recall score using None as sample_weights, None as labels, 1 as pos_label, binary as average (higher is better) : + - Score on train : 1.0 + - Score on test : 0.512195121951 + For ROS AUC score using None as sample_weights, micro as average (higher is better) : + - Score on train : 1.0 + - Score on test : 0.552015928323 + For Zero one loss using None as sample_weights (lower is better) : + - Score on train : 0.0 + - Score on test : 0.444444444444 + + + Classification took 0:00:00 \ No newline at end of file diff --git a/Code/MonoMutliViewClassifiers/Results/20160906-112157Results-DecisionTree-Non-Oui-learnRate0.7-Fake.txt b/Code/MonoMutliViewClassifiers/Results/20160906-112157Results-DecisionTree-Non-Oui-learnRate0.7-Fake.txt new file mode 100644 index 0000000000000000000000000000000000000000..c42bb400ae8a8fdedb841cb3c754293d696319cb --- /dev/null +++ b/Code/MonoMutliViewClassifiers/Results/20160906-112157Results-DecisionTree-Non-Oui-learnRate0.7-Fake.txt @@ -0,0 +1,54 @@ +Classification on Fake database for View2 with DecisionTree + +accuracy_score on train : 0.995238095238 +accuracy_score on test : 0.533333333333 + +Database configuration : + - Database name : Fake + - View name : View2 View shape : (300, 8) + - Learning Rate : 0.7 + - Labels used : Non, Oui + - Number of cross validation folds : 5 + +Classifier configuration : + - Decision Tree with max_depth : 15 + - Executed on 1 core(s) + - Got configuration using randomized search with 1 iterations + + + For Accuracy score using None as sample_weights (higher is better) : + - Score on train : 0.995238095238 + - Score on test : 0.533333333333 + For F1 score using None as sample_weights, None as labels, 1 as pos_label, micro as average (higher is better) : + - Score on train : 0.995305164319 + - Score on test : 0.511627906977 + For F-beta score using None as sample_weights, None as labels, 1 as pos_label, micro as average, 1.0 as beta (higher is better) : + - Score on train : 0.995305164319 + - Score on test : 0.511627906977 + For Hamming loss using None as classes (lower is better) : + - Score on train : 0.0047619047619 + - Score on test : 0.466666666667 + For Jaccard similarity score using None as sample_weights (higher is better) : + - Score on train : 0.995238095238 + - Score on test : 0.533333333333 + For Log loss using None as sample_weights, 1e-15 as eps (lower is better) : + - Score on train : nan + - Score on test : nan + For Matthews correlation coefficient (higher is better) : + - Score on train : 0.990519401324 + - Score on test : 0.066931612238 + For Precision score using None as sample_weights, None as labels, 1 as pos_label, binary as average (higher is better) : + - Score on train : 0.990654205607 + - Score on test : 0.488888888889 + For Recall score using None as sample_weights, None as labels, 1 as pos_label, binary as average (higher is better) : + - Score on train : 1.0 + - Score on test : 0.536585365854 + For ROS AUC score using None as sample_weights, micro as average (higher is better) : + - Score on train : 0.995192307692 + - Score on test : 0.533598805376 + For Zero one loss using None as sample_weights (lower is better) : + - Score on train : 0.0047619047619 + - Score on test : 0.466666666667 + + + Classification took 0:00:00 \ No newline at end of file diff --git a/Code/MonoMutliViewClassifiers/Results/20160906-112157Results-RandomForest-Non-Oui-learnRate0.7-Fake.txt b/Code/MonoMutliViewClassifiers/Results/20160906-112157Results-RandomForest-Non-Oui-learnRate0.7-Fake.txt new file mode 100644 index 0000000000000000000000000000000000000000..7bf769f1bc0c2c156f175b1fe70658f88c9b6906 --- /dev/null +++ b/Code/MonoMutliViewClassifiers/Results/20160906-112157Results-RandomForest-Non-Oui-learnRate0.7-Fake.txt @@ -0,0 +1,54 @@ +Classification on Fake database for View1 with RandomForest + +accuracy_score on train : 0.97619047619 +accuracy_score on test : 0.522222222222 + +Database configuration : + - Database name : Fake + - View name : View1 View shape : (300, 18) + - Learning Rate : 0.7 + - Labels used : Non, Oui + - Number of cross validation folds : 5 + +Classifier configuration : + - Random Forest with num_esimators : 7, max_depth : 15 + - Executed on 1 core(s) + - Got configuration using randomized search with 1 iterations + + + For Accuracy score using None as sample_weights (higher is better) : + - Score on train : 0.97619047619 + - Score on test : 0.522222222222 + For F1 score using None as sample_weights, None as labels, 1 as pos_label, micro as average (higher is better) : + - Score on train : 0.976525821596 + - Score on test : 0.516853932584 + For F-beta score using None as sample_weights, None as labels, 1 as pos_label, micro as average, 1.0 as beta (higher is better) : + - Score on train : 0.976525821596 + - Score on test : 0.516853932584 + For Hamming loss using None as classes (lower is better) : + - Score on train : 0.0238095238095 + - Score on test : 0.477777777778 + For Jaccard similarity score using None as sample_weights (higher is better) : + - Score on train : 0.97619047619 + - Score on test : 0.522222222222 + For Log loss using None as sample_weights, 1e-15 as eps (lower is better) : + - Score on train : nan + - Score on test : nan + For Matthews correlation coefficient (higher is better) : + - Score on train : 0.952415522541 + - Score on test : 0.0506833064614 + For Precision score using None as sample_weights, None as labels, 1 as pos_label, binary as average (higher is better) : + - Score on train : 0.971962616822 + - Score on test : 0.479166666667 + For Recall score using None as sample_weights, None as labels, 1 as pos_label, binary as average (higher is better) : + - Score on train : 0.981132075472 + - Score on test : 0.560975609756 + For ROS AUC score using None as sample_weights, micro as average (higher is better) : + - Score on train : 0.976142960813 + - Score on test : 0.525385764062 + For Zero one loss using None as sample_weights (lower is better) : + - Score on train : 0.0238095238095 + - Score on test : 0.477777777778 + + + Classification took 0:00:00 \ No newline at end of file diff --git a/Code/MonoMutliViewClassifiers/Results/20160906-112157Results-SGD-Non-Oui-learnRate0.7-Fake.txt b/Code/MonoMutliViewClassifiers/Results/20160906-112157Results-SGD-Non-Oui-learnRate0.7-Fake.txt new file mode 100644 index 0000000000000000000000000000000000000000..45ae81a4f03834da811af554902b0ff07b317177 --- /dev/null +++ b/Code/MonoMutliViewClassifiers/Results/20160906-112157Results-SGD-Non-Oui-learnRate0.7-Fake.txt @@ -0,0 +1,54 @@ +Classification on Fake database for View1 with SGD + +accuracy_score on train : 0.566666666667 +accuracy_score on test : 0.522222222222 + +Database configuration : + - Database name : Fake + - View name : View1 View shape : (300, 18) + - Learning Rate : 0.7 + - Labels used : Non, Oui + - Number of cross validation folds : 5 + +Classifier configuration : + - SGDClassifier with loss : modified_huber, penalty : l2 + - Executed on 1 core(s) + - Got configuration using randomized search with 1 iterations + + + For Accuracy score using None as sample_weights (higher is better) : + - Score on train : 0.566666666667 + - Score on test : 0.522222222222 + For F1 score using None as sample_weights, None as labels, 1 as pos_label, micro as average (higher is better) : + - Score on train : 0.564593301435 + - Score on test : 0.516853932584 + For F-beta score using None as sample_weights, None as labels, 1 as pos_label, micro as average, 1.0 as beta (higher is better) : + - Score on train : 0.564593301435 + - Score on test : 0.516853932584 + For Hamming loss using None as classes (lower is better) : + - Score on train : 0.433333333333 + - Score on test : 0.477777777778 + For Jaccard similarity score using None as sample_weights (higher is better) : + - Score on train : 0.566666666667 + - Score on test : 0.522222222222 + For Log loss using None as sample_weights, 1e-15 as eps (lower is better) : + - Score on train : nan + - Score on test : nan + For Matthews correlation coefficient (higher is better) : + - Score on train : 0.133545022783 + - Score on test : 0.0506833064614 + For Precision score using None as sample_weights, None as labels, 1 as pos_label, binary as average (higher is better) : + - Score on train : 0.572815533981 + - Score on test : 0.479166666667 + For Recall score using None as sample_weights, None as labels, 1 as pos_label, binary as average (higher is better) : + - Score on train : 0.556603773585 + - Score on test : 0.560975609756 + For ROS AUC score using None as sample_weights, micro as average (higher is better) : + - Score on train : 0.566763425254 + - Score on test : 0.525385764062 + For Zero one loss using None as sample_weights (lower is better) : + - Score on train : 0.433333333333 + - Score on test : 0.477777777778 + + + Classification took 0:00:00 \ No newline at end of file diff --git a/Code/MonoMutliViewClassifiers/Results/20160906-112157Results-SVMLinear-Non-Oui-learnRate0.7-Fake.txt b/Code/MonoMutliViewClassifiers/Results/20160906-112157Results-SVMLinear-Non-Oui-learnRate0.7-Fake.txt new file mode 100644 index 0000000000000000000000000000000000000000..395792e34ea36d0f71040c5bd654478caf5c099b --- /dev/null +++ b/Code/MonoMutliViewClassifiers/Results/20160906-112157Results-SVMLinear-Non-Oui-learnRate0.7-Fake.txt @@ -0,0 +1,54 @@ +Classification on Fake database for View1 with SVMLinear + +accuracy_score on train : 0.528571428571 +accuracy_score on test : 0.533333333333 + +Database configuration : + - Database name : Fake + - View name : View1 View shape : (300, 18) + - Learning Rate : 0.7 + - Labels used : Non, Oui + - Number of cross validation folds : 5 + +Classifier configuration : + - SVM Linear with C : 2991 + - Executed on 1 core(s) + - Got configuration using randomized search with 1 iterations + + + For Accuracy score using None as sample_weights (higher is better) : + - Score on train : 0.528571428571 + - Score on test : 0.533333333333 + For F1 score using None as sample_weights, None as labels, 1 as pos_label, micro as average (higher is better) : + - Score on train : 0.535211267606 + - Score on test : 0.475 + For F-beta score using None as sample_weights, None as labels, 1 as pos_label, micro as average, 1.0 as beta (higher is better) : + - Score on train : 0.535211267606 + - Score on test : 0.475 + For Hamming loss using None as classes (lower is better) : + - Score on train : 0.471428571429 + - Score on test : 0.466666666667 + For Jaccard similarity score using None as sample_weights (higher is better) : + - Score on train : 0.528571428571 + - Score on test : 0.533333333333 + For Log loss using None as sample_weights, 1e-15 as eps (lower is better) : + - Score on train : nan + - Score on test : nan + For Matthews correlation coefficient (higher is better) : + - Score on train : 0.0569743711331 + - Score on test : 0.0555284586866 + For Precision score using None as sample_weights, None as labels, 1 as pos_label, binary as average (higher is better) : + - Score on train : 0.532710280374 + - Score on test : 0.487179487179 + For Recall score using None as sample_weights, None as labels, 1 as pos_label, binary as average (higher is better) : + - Score on train : 0.537735849057 + - Score on test : 0.463414634146 + For ROS AUC score using None as sample_weights, micro as average (higher is better) : + - Score on train : 0.528483309144 + - Score on test : 0.52762568442 + For Zero one loss using None as sample_weights (lower is better) : + - Score on train : 0.471428571429 + - Score on test : 0.466666666667 + + + Classification took 0:00:00 \ No newline at end of file diff --git a/Code/MonoMutliViewClassifiers/Results/20160906-112157Results-SVMPoly-Non-Oui-learnRate0.7-Fake.txt b/Code/MonoMutliViewClassifiers/Results/20160906-112157Results-SVMPoly-Non-Oui-learnRate0.7-Fake.txt new file mode 100644 index 0000000000000000000000000000000000000000..96ee48b4709eca185b46819a817e5c8f27db9696 --- /dev/null +++ b/Code/MonoMutliViewClassifiers/Results/20160906-112157Results-SVMPoly-Non-Oui-learnRate0.7-Fake.txt @@ -0,0 +1,54 @@ +Classification on Fake database for View1 with SVMPoly + +accuracy_score on train : 1.0 +accuracy_score on test : 0.444444444444 + +Database configuration : + - Database name : Fake + - View name : View1 View shape : (300, 18) + - Learning Rate : 0.7 + - Labels used : Non, Oui + - Number of cross validation folds : 5 + +Classifier configuration : + - SVM Linear with C : 2991 + - Executed on 1 core(s) + - Got configuration using randomized search with 1 iterations + + + For Accuracy score using None as sample_weights (higher is better) : + - Score on train : 1.0 + - Score on test : 0.444444444444 + For F1 score using None as sample_weights, None as labels, 1 as pos_label, micro as average (higher is better) : + - Score on train : 1.0 + - Score on test : 0.561403508772 + For F-beta score using None as sample_weights, None as labels, 1 as pos_label, micro as average, 1.0 as beta (higher is better) : + - Score on train : 1.0 + - Score on test : 0.561403508772 + For Hamming loss using None as classes (lower is better) : + - Score on train : 0.0 + - Score on test : 0.555555555556 + For Jaccard similarity score using None as sample_weights (higher is better) : + - Score on train : 1.0 + - Score on test : 0.444444444444 + For Log loss using None as sample_weights, 1e-15 as eps (lower is better) : + - Score on train : nan + - Score on test : nan + For Matthews correlation coefficient (higher is better) : + - Score on train : 1.0 + - Score on test : -0.0715653145323 + For Precision score using None as sample_weights, None as labels, 1 as pos_label, binary as average (higher is better) : + - Score on train : 1.0 + - Score on test : 0.438356164384 + For Recall score using None as sample_weights, None as labels, 1 as pos_label, binary as average (higher is better) : + - Score on train : 1.0 + - Score on test : 0.780487804878 + For ROS AUC score using None as sample_weights, micro as average (higher is better) : + - Score on train : 1.0 + - Score on test : 0.4718765555 + For Zero one loss using None as sample_weights (lower is better) : + - Score on train : 0.0 + - Score on test : 0.555555555556 + + + Classification took 0:00:00 \ No newline at end of file diff --git a/Code/MonoMutliViewClassifiers/Results/20160906-112157Results-SVMRBF-Non-Oui-learnRate0.7-Fake.txt b/Code/MonoMutliViewClassifiers/Results/20160906-112157Results-SVMRBF-Non-Oui-learnRate0.7-Fake.txt new file mode 100644 index 0000000000000000000000000000000000000000..75cb5217f476620552e022b850221b11be87393f --- /dev/null +++ b/Code/MonoMutliViewClassifiers/Results/20160906-112157Results-SVMRBF-Non-Oui-learnRate0.7-Fake.txt @@ -0,0 +1,54 @@ +Classification on Fake database for View1 with SVMRBF + +accuracy_score on train : 1.0 +accuracy_score on test : 0.533333333333 + +Database configuration : + - Database name : Fake + - View name : View1 View shape : (300, 18) + - Learning Rate : 0.7 + - Labels used : Non, Oui + - Number of cross validation folds : 5 + +Classifier configuration : + - SVM Linear with C : 2991 + - Executed on 1 core(s) + - Got configuration using randomized search with 1 iterations + + + For Accuracy score using None as sample_weights (higher is better) : + - Score on train : 1.0 + - Score on test : 0.533333333333 + For F1 score using None as sample_weights, None as labels, 1 as pos_label, micro as average (higher is better) : + - Score on train : 1.0 + - Score on test : 0.511627906977 + For F-beta score using None as sample_weights, None as labels, 1 as pos_label, micro as average, 1.0 as beta (higher is better) : + - Score on train : 1.0 + - Score on test : 0.511627906977 + For Hamming loss using None as classes (lower is better) : + - Score on train : 0.0 + - Score on test : 0.466666666667 + For Jaccard similarity score using None as sample_weights (higher is better) : + - Score on train : 1.0 + - Score on test : 0.533333333333 + For Log loss using None as sample_weights, 1e-15 as eps (lower is better) : + - Score on train : nan + - Score on test : nan + For Matthews correlation coefficient (higher is better) : + - Score on train : 1.0 + - Score on test : 0.066931612238 + For Precision score using None as sample_weights, None as labels, 1 as pos_label, binary as average (higher is better) : + - Score on train : 1.0 + - Score on test : 0.488888888889 + For Recall score using None as sample_weights, None as labels, 1 as pos_label, binary as average (higher is better) : + - Score on train : 1.0 + - Score on test : 0.536585365854 + For ROS AUC score using None as sample_weights, micro as average (higher is better) : + - Score on train : 1.0 + - Score on test : 0.533598805376 + For Zero one loss using None as sample_weights (lower is better) : + - Score on train : 0.0 + - Score on test : 0.466666666667 + + + Classification took 0:00:00 \ No newline at end of file diff --git a/Code/MonoMutliViewClassifiers/Results/20160906-112158Results-Adaboost-Non-Oui-learnRate0.7-Fake.txt b/Code/MonoMutliViewClassifiers/Results/20160906-112158Results-Adaboost-Non-Oui-learnRate0.7-Fake.txt new file mode 100644 index 0000000000000000000000000000000000000000..92282ac5d982afa7b67e798392101b6eab92cabd --- /dev/null +++ b/Code/MonoMutliViewClassifiers/Results/20160906-112158Results-Adaboost-Non-Oui-learnRate0.7-Fake.txt @@ -0,0 +1,57 @@ +Classification on Fake database for View3 with Adaboost + +accuracy_score on train : 1.0 +accuracy_score on test : 0.444444444444 + +Database configuration : + - Database name : Fake + - View name : View3 View shape : (300, 6) + - Learning Rate : 0.7 + - Labels used : Non, Oui + - Number of cross validation folds : 5 + +Classifier configuration : + - Adaboost with num_esimators : 7, base_estimators : DecisionTreeClassifier(class_weight=None, criterion='gini', max_depth=None, + max_features=None, max_leaf_nodes=None, min_samples_leaf=1, + min_samples_split=2, min_weight_fraction_leaf=0.0, + presort=False, random_state=None, splitter='best') + - Executed on 1 core(s) + - Got configuration using randomized search with 1 iterations + + + For Accuracy score using None as sample_weights (higher is better) : + - Score on train : 1.0 + - Score on test : 0.444444444444 + For F1 score using None as sample_weights, None as labels, 1 as pos_label, micro as average (higher is better) : + - Score on train : 1.0 + - Score on test : 0.404761904762 + For F-beta score using None as sample_weights, None as labels, 1 as pos_label, micro as average, 1.0 as beta (higher is better) : + - Score on train : 1.0 + - Score on test : 0.404761904762 + For Hamming loss using None as classes (lower is better) : + - Score on train : 0.0 + - Score on test : 0.555555555556 + For Jaccard similarity score using None as sample_weights (higher is better) : + - Score on train : 1.0 + - Score on test : 0.444444444444 + For Log loss using None as sample_weights, 1e-15 as eps (lower is better) : + - Score on train : nan + - Score on test : nan + For Matthews correlation coefficient (higher is better) : + - Score on train : 1.0 + - Score on test : -0.115633266975 + For Precision score using None as sample_weights, None as labels, 1 as pos_label, binary as average (higher is better) : + - Score on train : 1.0 + - Score on test : 0.395348837209 + For Recall score using None as sample_weights, None as labels, 1 as pos_label, binary as average (higher is better) : + - Score on train : 1.0 + - Score on test : 0.414634146341 + For ROS AUC score using None as sample_weights, micro as average (higher is better) : + - Score on train : 1.0 + - Score on test : 0.442010950722 + For Zero one loss using None as sample_weights (lower is better) : + - Score on train : 0.0 + - Score on test : 0.555555555556 + + + Classification took 0:00:00 \ No newline at end of file diff --git a/Code/MonoMutliViewClassifiers/Results/20160906-112158Results-DecisionTree-Non-Oui-learnRate0.7-Fake.txt b/Code/MonoMutliViewClassifiers/Results/20160906-112158Results-DecisionTree-Non-Oui-learnRate0.7-Fake.txt new file mode 100644 index 0000000000000000000000000000000000000000..ed81b59a61d0887a2038565f79039b2a5ebb6c36 --- /dev/null +++ b/Code/MonoMutliViewClassifiers/Results/20160906-112158Results-DecisionTree-Non-Oui-learnRate0.7-Fake.txt @@ -0,0 +1,54 @@ +Classification on Fake database for View3 with DecisionTree + +accuracy_score on train : 1.0 +accuracy_score on test : 0.466666666667 + +Database configuration : + - Database name : Fake + - View name : View3 View shape : (300, 6) + - Learning Rate : 0.7 + - Labels used : Non, Oui + - Number of cross validation folds : 5 + +Classifier configuration : + - Decision Tree with max_depth : 15 + - Executed on 1 core(s) + - Got configuration using randomized search with 1 iterations + + + For Accuracy score using None as sample_weights (higher is better) : + - Score on train : 1.0 + - Score on test : 0.466666666667 + For F1 score using None as sample_weights, None as labels, 1 as pos_label, micro as average (higher is better) : + - Score on train : 1.0 + - Score on test : 0.428571428571 + For F-beta score using None as sample_weights, None as labels, 1 as pos_label, micro as average, 1.0 as beta (higher is better) : + - Score on train : 1.0 + - Score on test : 0.428571428571 + For Hamming loss using None as classes (lower is better) : + - Score on train : 0.0 + - Score on test : 0.533333333333 + For Jaccard similarity score using None as sample_weights (higher is better) : + - Score on train : 1.0 + - Score on test : 0.466666666667 + For Log loss using None as sample_weights, 1e-15 as eps (lower is better) : + - Score on train : nan + - Score on test : nan + For Matthews correlation coefficient (higher is better) : + - Score on train : 1.0 + - Score on test : -0.0709680565554 + For Precision score using None as sample_weights, None as labels, 1 as pos_label, binary as average (higher is better) : + - Score on train : 1.0 + - Score on test : 0.418604651163 + For Recall score using None as sample_weights, None as labels, 1 as pos_label, binary as average (higher is better) : + - Score on train : 1.0 + - Score on test : 0.439024390244 + For ROS AUC score using None as sample_weights, micro as average (higher is better) : + - Score on train : 1.0 + - Score on test : 0.464410154306 + For Zero one loss using None as sample_weights (lower is better) : + - Score on train : 0.0 + - Score on test : 0.533333333333 + + + Classification took 0:00:00 \ No newline at end of file diff --git a/Code/MonoMutliViewClassifiers/Results/20160906-112158Results-KNN-Non-Oui-learnRate0.7-Fake.txt b/Code/MonoMutliViewClassifiers/Results/20160906-112158Results-KNN-Non-Oui-learnRate0.7-Fake.txt new file mode 100644 index 0000000000000000000000000000000000000000..bddd0960441ccf4c67f24b028906a144474cd1c2 --- /dev/null +++ b/Code/MonoMutliViewClassifiers/Results/20160906-112158Results-KNN-Non-Oui-learnRate0.7-Fake.txt @@ -0,0 +1,54 @@ +Classification on Fake database for View2 with KNN + +accuracy_score on train : 0.552380952381 +accuracy_score on test : 0.444444444444 + +Database configuration : + - Database name : Fake + - View name : View2 View shape : (300, 8) + - Learning Rate : 0.7 + - Labels used : Non, Oui + - Number of cross validation folds : 5 + +Classifier configuration : + - K nearest Neighbors with n_neighbors: 47 + - Executed on 1 core(s) + - Got configuration using randomized search with 1 iterations + + + For Accuracy score using None as sample_weights (higher is better) : + - Score on train : 0.552380952381 + - Score on test : 0.444444444444 + For F1 score using None as sample_weights, None as labels, 1 as pos_label, micro as average (higher is better) : + - Score on train : 0.477777777778 + - Score on test : 0.342105263158 + For F-beta score using None as sample_weights, None as labels, 1 as pos_label, micro as average, 1.0 as beta (higher is better) : + - Score on train : 0.477777777778 + - Score on test : 0.342105263158 + For Hamming loss using None as classes (lower is better) : + - Score on train : 0.447619047619 + - Score on test : 0.555555555556 + For Jaccard similarity score using None as sample_weights (higher is better) : + - Score on train : 0.552380952381 + - Score on test : 0.444444444444 + For Log loss using None as sample_weights, 1e-15 as eps (lower is better) : + - Score on train : nan + - Score on test : nan + For Matthews correlation coefficient (higher is better) : + - Score on train : 0.112597765671 + - Score on test : -0.134753650348 + For Precision score using None as sample_weights, None as labels, 1 as pos_label, binary as average (higher is better) : + - Score on train : 0.581081081081 + - Score on test : 0.371428571429 + For Recall score using None as sample_weights, None as labels, 1 as pos_label, binary as average (higher is better) : + - Score on train : 0.405660377358 + - Score on test : 0.317073170732 + For ROS AUC score using None as sample_weights, micro as average (higher is better) : + - Score on train : 0.553791727141 + - Score on test : 0.434046789447 + For Zero one loss using None as sample_weights (lower is better) : + - Score on train : 0.447619047619 + - Score on test : 0.555555555556 + + + Classification took 0:00:00 \ No newline at end of file diff --git a/Code/MonoMutliViewClassifiers/Results/20160906-112158Results-RandomForest-Non-Oui-learnRate0.7-Fake.txt b/Code/MonoMutliViewClassifiers/Results/20160906-112158Results-RandomForest-Non-Oui-learnRate0.7-Fake.txt new file mode 100644 index 0000000000000000000000000000000000000000..9d35d1a6da182138e6ae1ad331a4e2d7316dbb03 --- /dev/null +++ b/Code/MonoMutliViewClassifiers/Results/20160906-112158Results-RandomForest-Non-Oui-learnRate0.7-Fake.txt @@ -0,0 +1,54 @@ +Classification on Fake database for View2 with RandomForest + +accuracy_score on train : 0.966666666667 +accuracy_score on test : 0.588888888889 + +Database configuration : + - Database name : Fake + - View name : View2 View shape : (300, 8) + - Learning Rate : 0.7 + - Labels used : Non, Oui + - Number of cross validation folds : 5 + +Classifier configuration : + - Random Forest with num_esimators : 7, max_depth : 15 + - Executed on 1 core(s) + - Got configuration using randomized search with 1 iterations + + + For Accuracy score using None as sample_weights (higher is better) : + - Score on train : 0.966666666667 + - Score on test : 0.588888888889 + For F1 score using None as sample_weights, None as labels, 1 as pos_label, micro as average (higher is better) : + - Score on train : 0.96682464455 + - Score on test : 0.53164556962 + For F-beta score using None as sample_weights, None as labels, 1 as pos_label, micro as average, 1.0 as beta (higher is better) : + - Score on train : 0.96682464455 + - Score on test : 0.53164556962 + For Hamming loss using None as classes (lower is better) : + - Score on train : 0.0333333333333 + - Score on test : 0.411111111111 + For Jaccard similarity score using None as sample_weights (higher is better) : + - Score on train : 0.966666666667 + - Score on test : 0.588888888889 + For Log loss using None as sample_weights, 1e-15 as eps (lower is better) : + - Score on train : nan + - Score on test : nan + For Matthews correlation coefficient (higher is better) : + - Score on train : 0.933375664255 + - Score on test : 0.166630556676 + For Precision score using None as sample_weights, None as labels, 1 as pos_label, binary as average (higher is better) : + - Score on train : 0.971428571429 + - Score on test : 0.552631578947 + For Recall score using None as sample_weights, None as labels, 1 as pos_label, binary as average (higher is better) : + - Score on train : 0.962264150943 + - Score on test : 0.512195121951 + For ROS AUC score using None as sample_weights, micro as average (higher is better) : + - Score on train : 0.966708998549 + - Score on test : 0.582628173221 + For Zero one loss using None as sample_weights (lower is better) : + - Score on train : 0.0333333333333 + - Score on test : 0.411111111111 + + + Classification took 0:00:00 \ No newline at end of file diff --git a/Code/MonoMutliViewClassifiers/Results/20160906-112158Results-SGD-Non-Oui-learnRate0.7-Fake.txt b/Code/MonoMutliViewClassifiers/Results/20160906-112158Results-SGD-Non-Oui-learnRate0.7-Fake.txt new file mode 100644 index 0000000000000000000000000000000000000000..98fcb2e6715737097b534f2d5a4e9c1553160301 --- /dev/null +++ b/Code/MonoMutliViewClassifiers/Results/20160906-112158Results-SGD-Non-Oui-learnRate0.7-Fake.txt @@ -0,0 +1,54 @@ +Classification on Fake database for View2 with SGD + +accuracy_score on train : 0.561904761905 +accuracy_score on test : 0.5 + +Database configuration : + - Database name : Fake + - View name : View2 View shape : (300, 8) + - Learning Rate : 0.7 + - Labels used : Non, Oui + - Number of cross validation folds : 5 + +Classifier configuration : + - SGDClassifier with loss : modified_huber, penalty : l2 + - Executed on 1 core(s) + - Got configuration using randomized search with 1 iterations + + + For Accuracy score using None as sample_weights (higher is better) : + - Score on train : 0.561904761905 + - Score on test : 0.5 + For F1 score using None as sample_weights, None as labels, 1 as pos_label, micro as average (higher is better) : + - Score on train : 0.544554455446 + - Score on test : 0.470588235294 + For F-beta score using None as sample_weights, None as labels, 1 as pos_label, micro as average, 1.0 as beta (higher is better) : + - Score on train : 0.544554455446 + - Score on test : 0.470588235294 + For Hamming loss using None as classes (lower is better) : + - Score on train : 0.438095238095 + - Score on test : 0.5 + For Jaccard similarity score using None as sample_weights (higher is better) : + - Score on train : 0.561904761905 + - Score on test : 0.5 + For Log loss using None as sample_weights, 1e-15 as eps (lower is better) : + - Score on train : nan + - Score on test : nan + For Matthews correlation coefficient (higher is better) : + - Score on train : 0.125091870983 + - Score on test : -0.00198364873142 + For Precision score using None as sample_weights, None as labels, 1 as pos_label, binary as average (higher is better) : + - Score on train : 0.572916666667 + - Score on test : 0.454545454545 + For Recall score using None as sample_weights, None as labels, 1 as pos_label, binary as average (higher is better) : + - Score on train : 0.518867924528 + - Score on test : 0.487804878049 + For ROS AUC score using None as sample_weights, micro as average (higher is better) : + - Score on train : 0.562318577649 + - Score on test : 0.499004479841 + For Zero one loss using None as sample_weights (lower is better) : + - Score on train : 0.438095238095 + - Score on test : 0.5 + + + Classification took 0:00:00 \ No newline at end of file diff --git a/Code/MonoMutliViewClassifiers/Results/20160906-112158Results-SVMLinear-Non-Oui-learnRate0.7-Fake.txt b/Code/MonoMutliViewClassifiers/Results/20160906-112158Results-SVMLinear-Non-Oui-learnRate0.7-Fake.txt new file mode 100644 index 0000000000000000000000000000000000000000..4471f09ed96536e0167cd416edb89fe3dff7ef2b --- /dev/null +++ b/Code/MonoMutliViewClassifiers/Results/20160906-112158Results-SVMLinear-Non-Oui-learnRate0.7-Fake.txt @@ -0,0 +1,54 @@ +Classification on Fake database for View2 with SVMLinear + +accuracy_score on train : 0.495238095238 +accuracy_score on test : 0.633333333333 + +Database configuration : + - Database name : Fake + - View name : View2 View shape : (300, 8) + - Learning Rate : 0.7 + - Labels used : Non, Oui + - Number of cross validation folds : 5 + +Classifier configuration : + - SVM Linear with C : 2991 + - Executed on 1 core(s) + - Got configuration using randomized search with 1 iterations + + + For Accuracy score using None as sample_weights (higher is better) : + - Score on train : 0.495238095238 + - Score on test : 0.633333333333 + For F1 score using None as sample_weights, None as labels, 1 as pos_label, micro as average (higher is better) : + - Score on train : 0.51376146789 + - Score on test : 0.637362637363 + For F-beta score using None as sample_weights, None as labels, 1 as pos_label, micro as average, 1.0 as beta (higher is better) : + - Score on train : 0.51376146789 + - Score on test : 0.637362637363 + For Hamming loss using None as classes (lower is better) : + - Score on train : 0.504761904762 + - Score on test : 0.366666666667 + For Jaccard similarity score using None as sample_weights (higher is better) : + - Score on train : 0.495238095238 + - Score on test : 0.633333333333 + For Log loss using None as sample_weights, 1e-15 as eps (lower is better) : + - Score on train : nan + - Score on test : nan + For Matthews correlation coefficient (higher is better) : + - Score on train : -0.0101818424163 + - Score on test : 0.279372118308 + For Precision score using None as sample_weights, None as labels, 1 as pos_label, binary as average (higher is better) : + - Score on train : 0.5 + - Score on test : 0.58 + For Recall score using None as sample_weights, None as labels, 1 as pos_label, binary as average (higher is better) : + - Score on train : 0.528301886792 + - Score on test : 0.707317073171 + For ROS AUC score using None as sample_weights, micro as average (higher is better) : + - Score on train : 0.494920174165 + - Score on test : 0.6393728223 + For Zero one loss using None as sample_weights (lower is better) : + - Score on train : 0.504761904762 + - Score on test : 0.366666666667 + + + Classification took 0:00:00 \ No newline at end of file diff --git a/Code/MonoMutliViewClassifiers/Results/20160906-112158Results-SVMPoly-Non-Oui-learnRate0.7-Fake.txt b/Code/MonoMutliViewClassifiers/Results/20160906-112158Results-SVMPoly-Non-Oui-learnRate0.7-Fake.txt new file mode 100644 index 0000000000000000000000000000000000000000..25f841d0b5e3d23af058cf951b334597f5e1f51a --- /dev/null +++ b/Code/MonoMutliViewClassifiers/Results/20160906-112158Results-SVMPoly-Non-Oui-learnRate0.7-Fake.txt @@ -0,0 +1,54 @@ +Classification on Fake database for View2 with SVMPoly + +accuracy_score on train : 0.995238095238 +accuracy_score on test : 0.444444444444 + +Database configuration : + - Database name : Fake + - View name : View2 View shape : (300, 8) + - Learning Rate : 0.7 + - Labels used : Non, Oui + - Number of cross validation folds : 5 + +Classifier configuration : + - SVM Linear with C : 2991 + - Executed on 1 core(s) + - Got configuration using randomized search with 1 iterations + + + For Accuracy score using None as sample_weights (higher is better) : + - Score on train : 0.995238095238 + - Score on test : 0.444444444444 + For F1 score using None as sample_weights, None as labels, 1 as pos_label, micro as average (higher is better) : + - Score on train : 0.995260663507 + - Score on test : 0.479166666667 + For F-beta score using None as sample_weights, None as labels, 1 as pos_label, micro as average, 1.0 as beta (higher is better) : + - Score on train : 0.995260663507 + - Score on test : 0.479166666667 + For Hamming loss using None as classes (lower is better) : + - Score on train : 0.0047619047619 + - Score on test : 0.555555555556 + For Jaccard similarity score using None as sample_weights (higher is better) : + - Score on train : 0.995238095238 + - Score on test : 0.444444444444 + For Log loss using None as sample_weights, 1e-15 as eps (lower is better) : + - Score on train : nan + - Score on test : nan + For Matthews correlation coefficient (higher is better) : + - Score on train : 0.990521113087 + - Score on test : -0.0940733030728 + For Precision score using None as sample_weights, None as labels, 1 as pos_label, binary as average (higher is better) : + - Score on train : 1.0 + - Score on test : 0.418181818182 + For Recall score using None as sample_weights, None as labels, 1 as pos_label, binary as average (higher is better) : + - Score on train : 0.990566037736 + - Score on test : 0.560975609756 + For ROS AUC score using None as sample_weights, micro as average (higher is better) : + - Score on train : 0.995283018868 + - Score on test : 0.453957192633 + For Zero one loss using None as sample_weights (lower is better) : + - Score on train : 0.0047619047619 + - Score on test : 0.555555555556 + + + Classification took 0:00:00 \ No newline at end of file diff --git a/Code/MonoMutliViewClassifiers/Results/20160906-112158Results-SVMRBF-Non-Oui-learnRate0.7-Fake.txt b/Code/MonoMutliViewClassifiers/Results/20160906-112158Results-SVMRBF-Non-Oui-learnRate0.7-Fake.txt new file mode 100644 index 0000000000000000000000000000000000000000..70e52533fb7bce3e44528ba58ac0135e2e7730c8 --- /dev/null +++ b/Code/MonoMutliViewClassifiers/Results/20160906-112158Results-SVMRBF-Non-Oui-learnRate0.7-Fake.txt @@ -0,0 +1,54 @@ +Classification on Fake database for View2 with SVMRBF + +accuracy_score on train : 1.0 +accuracy_score on test : 0.466666666667 + +Database configuration : + - Database name : Fake + - View name : View2 View shape : (300, 8) + - Learning Rate : 0.7 + - Labels used : Non, Oui + - Number of cross validation folds : 5 + +Classifier configuration : + - SVM Linear with C : 2991 + - Executed on 1 core(s) + - Got configuration using randomized search with 1 iterations + + + For Accuracy score using None as sample_weights (higher is better) : + - Score on train : 1.0 + - Score on test : 0.466666666667 + For F1 score using None as sample_weights, None as labels, 1 as pos_label, micro as average (higher is better) : + - Score on train : 1.0 + - Score on test : 0.466666666667 + For F-beta score using None as sample_weights, None as labels, 1 as pos_label, micro as average, 1.0 as beta (higher is better) : + - Score on train : 1.0 + - Score on test : 0.466666666667 + For Hamming loss using None as classes (lower is better) : + - Score on train : 0.0 + - Score on test : 0.533333333333 + For Jaccard similarity score using None as sample_weights (higher is better) : + - Score on train : 1.0 + - Score on test : 0.466666666667 + For Log loss using None as sample_weights, 1e-15 as eps (lower is better) : + - Score on train : nan + - Score on test : nan + For Matthews correlation coefficient (higher is better) : + - Score on train : 1.0 + - Score on test : -0.0592334494774 + For Precision score using None as sample_weights, None as labels, 1 as pos_label, binary as average (higher is better) : + - Score on train : 1.0 + - Score on test : 0.428571428571 + For Recall score using None as sample_weights, None as labels, 1 as pos_label, binary as average (higher is better) : + - Score on train : 1.0 + - Score on test : 0.512195121951 + For ROS AUC score using None as sample_weights, micro as average (higher is better) : + - Score on train : 1.0 + - Score on test : 0.470383275261 + For Zero one loss using None as sample_weights (lower is better) : + - Score on train : 0.0 + - Score on test : 0.533333333333 + + + Classification took 0:00:00 \ No newline at end of file diff --git a/Code/MonoMutliViewClassifiers/Results/20160906-112159Results-KNN-Non-Oui-learnRate0.7-Fake.txt b/Code/MonoMutliViewClassifiers/Results/20160906-112159Results-KNN-Non-Oui-learnRate0.7-Fake.txt new file mode 100644 index 0000000000000000000000000000000000000000..e57793d4a36e3f6ef7af1095b20a8028d4f664c8 --- /dev/null +++ b/Code/MonoMutliViewClassifiers/Results/20160906-112159Results-KNN-Non-Oui-learnRate0.7-Fake.txt @@ -0,0 +1,54 @@ +Classification on Fake database for View3 with KNN + +accuracy_score on train : 0.519047619048 +accuracy_score on test : 0.411111111111 + +Database configuration : + - Database name : Fake + - View name : View3 View shape : (300, 6) + - Learning Rate : 0.7 + - Labels used : Non, Oui + - Number of cross validation folds : 5 + +Classifier configuration : + - K nearest Neighbors with n_neighbors: 47 + - Executed on 1 core(s) + - Got configuration using randomized search with 1 iterations + + + For Accuracy score using None as sample_weights (higher is better) : + - Score on train : 0.519047619048 + - Score on test : 0.411111111111 + For F1 score using None as sample_weights, None as labels, 1 as pos_label, micro as average (higher is better) : + - Score on train : 0.562770562771 + - Score on test : 0.453608247423 + For F-beta score using None as sample_weights, None as labels, 1 as pos_label, micro as average, 1.0 as beta (higher is better) : + - Score on train : 0.562770562771 + - Score on test : 0.453608247423 + For Hamming loss using None as classes (lower is better) : + - Score on train : 0.480952380952 + - Score on test : 0.588888888889 + For Jaccard similarity score using None as sample_weights (higher is better) : + - Score on train : 0.519047619048 + - Score on test : 0.411111111111 + For Log loss using None as sample_weights, 1e-15 as eps (lower is better) : + - Score on train : nan + - Score on test : nan + For Matthews correlation coefficient (higher is better) : + - Score on train : 0.0369594857345 + - Score on test : -0.161571085301 + For Precision score using None as sample_weights, None as labels, 1 as pos_label, binary as average (higher is better) : + - Score on train : 0.52 + - Score on test : 0.392857142857 + For Recall score using None as sample_weights, None as labels, 1 as pos_label, binary as average (higher is better) : + - Score on train : 0.61320754717 + - Score on test : 0.536585365854 + For ROS AUC score using None as sample_weights, micro as average (higher is better) : + - Score on train : 0.518142235123 + - Score on test : 0.421353907417 + For Zero one loss using None as sample_weights (lower is better) : + - Score on train : 0.480952380952 + - Score on test : 0.588888888889 + + + Classification took 0:00:00 \ No newline at end of file diff --git a/Code/MonoMutliViewClassifiers/Results/20160906-112159Results-RandomForest-Non-Oui-learnRate0.7-Fake.txt b/Code/MonoMutliViewClassifiers/Results/20160906-112159Results-RandomForest-Non-Oui-learnRate0.7-Fake.txt new file mode 100644 index 0000000000000000000000000000000000000000..e01149648085dff1c6ff500035e0a1f24eec1e2d --- /dev/null +++ b/Code/MonoMutliViewClassifiers/Results/20160906-112159Results-RandomForest-Non-Oui-learnRate0.7-Fake.txt @@ -0,0 +1,54 @@ +Classification on Fake database for View3 with RandomForest + +accuracy_score on train : 0.97619047619 +accuracy_score on test : 0.488888888889 + +Database configuration : + - Database name : Fake + - View name : View3 View shape : (300, 6) + - Learning Rate : 0.7 + - Labels used : Non, Oui + - Number of cross validation folds : 5 + +Classifier configuration : + - Random Forest with num_esimators : 7, max_depth : 15 + - Executed on 1 core(s) + - Got configuration using randomized search with 1 iterations + + + For Accuracy score using None as sample_weights (higher is better) : + - Score on train : 0.97619047619 + - Score on test : 0.488888888889 + For F1 score using None as sample_weights, None as labels, 1 as pos_label, micro as average (higher is better) : + - Score on train : 0.976303317536 + - Score on test : 0.410256410256 + For F-beta score using None as sample_weights, None as labels, 1 as pos_label, micro as average, 1.0 as beta (higher is better) : + - Score on train : 0.976303317536 + - Score on test : 0.410256410256 + For Hamming loss using None as classes (lower is better) : + - Score on train : 0.0238095238095 + - Score on test : 0.511111111111 + For Jaccard similarity score using None as sample_weights (higher is better) : + - Score on train : 0.97619047619 + - Score on test : 0.488888888889 + For Log loss using None as sample_weights, 1e-15 as eps (lower is better) : + - Score on train : nan + - Score on test : nan + For Matthews correlation coefficient (higher is better) : + - Score on train : 0.952424147199 + - Score on test : -0.0387937676182 + For Precision score using None as sample_weights, None as labels, 1 as pos_label, binary as average (higher is better) : + - Score on train : 0.980952380952 + - Score on test : 0.432432432432 + For Recall score using None as sample_weights, None as labels, 1 as pos_label, binary as average (higher is better) : + - Score on train : 0.971698113208 + - Score on test : 0.390243902439 + For ROS AUC score using None as sample_weights, micro as average (higher is better) : + - Score on train : 0.976233671988 + - Score on test : 0.480836236934 + For Zero one loss using None as sample_weights (lower is better) : + - Score on train : 0.0238095238095 + - Score on test : 0.511111111111 + + + Classification took 0:00:00 \ No newline at end of file diff --git a/Code/MonoMutliViewClassifiers/Results/20160906-112159Results-SGD-Non-Oui-learnRate0.7-Fake.txt b/Code/MonoMutliViewClassifiers/Results/20160906-112159Results-SGD-Non-Oui-learnRate0.7-Fake.txt new file mode 100644 index 0000000000000000000000000000000000000000..08b2b5e9b92fa2691c9cf68a26275399640d2306 --- /dev/null +++ b/Code/MonoMutliViewClassifiers/Results/20160906-112159Results-SGD-Non-Oui-learnRate0.7-Fake.txt @@ -0,0 +1,54 @@ +Classification on Fake database for View3 with SGD + +accuracy_score on train : 0.542857142857 +accuracy_score on test : 0.466666666667 + +Database configuration : + - Database name : Fake + - View name : View3 View shape : (300, 6) + - Learning Rate : 0.7 + - Labels used : Non, Oui + - Number of cross validation folds : 5 + +Classifier configuration : + - SGDClassifier with loss : modified_huber, penalty : l2 + - Executed on 1 core(s) + - Got configuration using randomized search with 1 iterations + + + For Accuracy score using None as sample_weights (higher is better) : + - Score on train : 0.542857142857 + - Score on test : 0.466666666667 + For F1 score using None as sample_weights, None as labels, 1 as pos_label, micro as average (higher is better) : + - Score on train : 0.559633027523 + - Score on test : 0.478260869565 + For F-beta score using None as sample_weights, None as labels, 1 as pos_label, micro as average, 1.0 as beta (higher is better) : + - Score on train : 0.559633027523 + - Score on test : 0.478260869565 + For Hamming loss using None as classes (lower is better) : + - Score on train : 0.457142857143 + - Score on test : 0.533333333333 + For Jaccard similarity score using None as sample_weights (higher is better) : + - Score on train : 0.542857142857 + - Score on test : 0.466666666667 + For Log loss using None as sample_weights, 1e-15 as eps (lower is better) : + - Score on train : nan + - Score on test : nan + For Matthews correlation coefficient (higher is better) : + - Score on train : 0.0852729302366 + - Score on test : -0.0555284586866 + For Precision score using None as sample_weights, None as labels, 1 as pos_label, binary as average (higher is better) : + - Score on train : 0.544642857143 + - Score on test : 0.43137254902 + For Recall score using None as sample_weights, None as labels, 1 as pos_label, binary as average (higher is better) : + - Score on train : 0.575471698113 + - Score on test : 0.536585365854 + For ROS AUC score using None as sample_weights, micro as average (higher is better) : + - Score on train : 0.542543541364 + - Score on test : 0.47237431558 + For Zero one loss using None as sample_weights (lower is better) : + - Score on train : 0.457142857143 + - Score on test : 0.533333333333 + + + Classification took 0:00:00 \ No newline at end of file diff --git a/Code/MonoMutliViewClassifiers/Results/20160906-112159Results-SVMLinear-Non-Oui-learnRate0.7-Fake.txt b/Code/MonoMutliViewClassifiers/Results/20160906-112159Results-SVMLinear-Non-Oui-learnRate0.7-Fake.txt new file mode 100644 index 0000000000000000000000000000000000000000..75c3ebcffd6e56bb4cde5a45b8b7327951ad3ad2 --- /dev/null +++ b/Code/MonoMutliViewClassifiers/Results/20160906-112159Results-SVMLinear-Non-Oui-learnRate0.7-Fake.txt @@ -0,0 +1,54 @@ +Classification on Fake database for View3 with SVMLinear + +accuracy_score on train : 0.557142857143 +accuracy_score on test : 0.566666666667 + +Database configuration : + - Database name : Fake + - View name : View3 View shape : (300, 6) + - Learning Rate : 0.7 + - Labels used : Non, Oui + - Number of cross validation folds : 5 + +Classifier configuration : + - SVM Linear with C : 2991 + - Executed on 1 core(s) + - Got configuration using randomized search with 1 iterations + + + For Accuracy score using None as sample_weights (higher is better) : + - Score on train : 0.557142857143 + - Score on test : 0.566666666667 + For F1 score using None as sample_weights, None as labels, 1 as pos_label, micro as average (higher is better) : + - Score on train : 0.610878661088 + - Score on test : 0.571428571429 + For F-beta score using None as sample_weights, None as labels, 1 as pos_label, micro as average, 1.0 as beta (higher is better) : + - Score on train : 0.610878661088 + - Score on test : 0.571428571429 + For Hamming loss using None as classes (lower is better) : + - Score on train : 0.442857142857 + - Score on test : 0.433333333333 + For Jaccard similarity score using None as sample_weights (higher is better) : + - Score on train : 0.557142857143 + - Score on test : 0.566666666667 + For Log loss using None as sample_weights, 1e-15 as eps (lower is better) : + - Score on train : nan + - Score on test : nan + For Matthews correlation coefficient (higher is better) : + - Score on train : 0.11594977827 + - Score on test : 0.144674846981 + For Precision score using None as sample_weights, None as labels, 1 as pos_label, binary as average (higher is better) : + - Score on train : 0.548872180451 + - Score on test : 0.52 + For Recall score using None as sample_weights, None as labels, 1 as pos_label, binary as average (higher is better) : + - Score on train : 0.688679245283 + - Score on test : 0.634146341463 + For ROS AUC score using None as sample_weights, micro as average (higher is better) : + - Score on train : 0.55587808418 + - Score on test : 0.572175211548 + For Zero one loss using None as sample_weights (lower is better) : + - Score on train : 0.442857142857 + - Score on test : 0.433333333333 + + + Classification took 0:00:00 \ No newline at end of file diff --git a/Code/MonoMutliViewClassifiers/Results/20160906-161431-CMultiV-Benchmark-Methyl_MiRNA__RNASeq_Clinic-Fake-LOG.log b/Code/MonoMutliViewClassifiers/Results/20160906-161431-CMultiV-Benchmark-Methyl_MiRNA__RNASeq_Clinic-Fake-LOG.log new file mode 100644 index 0000000000000000000000000000000000000000..5d2fd1b2aa1e875c07cdef76f8f21a644659e684 --- /dev/null +++ b/Code/MonoMutliViewClassifiers/Results/20160906-161431-CMultiV-Benchmark-Methyl_MiRNA__RNASeq_Clinic-Fake-LOG.log @@ -0,0 +1,50 @@ +2016-09-06 16:14:31,063 DEBUG: Start: Creating 2 temporary datasets for multiprocessing +2016-09-06 16:14:31,063 WARNING: WARNING : /!\ This may use a lot of HDD storage space : 0.00012634375 Gbytes /!\ +2016-09-06 16:14:36,078 DEBUG: Start: Creating datasets for multiprocessing +2016-09-06 16:14:36,080 INFO: Start: Finding all available mono- & multiview algorithms +2016-09-06 16:14:36,142 INFO: ### Main Programm for Multiview Classification +2016-09-06 16:14:36,142 INFO: ### Classification - Database : Fake ; Views : Methyl, MiRNA_, RNASeq, Clinic ; Algorithm : Fusion ; Cores : 1 +2016-09-06 16:14:36,143 INFO: Info: Shape of View0 :(300, 9) +2016-09-06 16:14:36,143 INFO: ### Main Programm for Multiview Classification +2016-09-06 16:14:36,143 INFO: ### Classification - Database : Fake ; Views : Methyl, MiRNA_, RNASeq, Clinic ; Algorithm : Fusion ; Cores : 1 +2016-09-06 16:14:36,143 INFO: Info: Shape of View1 :(300, 13) +2016-09-06 16:14:36,144 INFO: Info: Shape of View0 :(300, 9) +2016-09-06 16:14:36,144 INFO: Info: Shape of View2 :(300, 12) +2016-09-06 16:14:36,144 INFO: Info: Shape of View1 :(300, 13) +2016-09-06 16:14:36,144 INFO: Info: Shape of View3 :(300, 12) +2016-09-06 16:14:36,145 INFO: Done: Read Database Files +2016-09-06 16:14:36,145 INFO: Start: Determine validation split for ratio 0.7 +2016-09-06 16:14:36,145 INFO: Info: Shape of View2 :(300, 12) +2016-09-06 16:14:36,145 INFO: Info: Shape of View3 :(300, 12) +2016-09-06 16:14:36,145 INFO: Done: Read Database Files +2016-09-06 16:14:36,145 INFO: Start: Determine validation split for ratio 0.7 +2016-09-06 16:14:36,150 INFO: Done: Determine validation split +2016-09-06 16:14:36,150 INFO: Start: Determine 5 folds +2016-09-06 16:14:36,151 INFO: Done: Determine validation split +2016-09-06 16:14:36,151 INFO: Start: Determine 5 folds +2016-09-06 16:14:36,159 INFO: Info: Length of Learning Sets: 169 +2016-09-06 16:14:36,159 INFO: Info: Length of Testing Sets: 42 +2016-09-06 16:14:36,159 INFO: Info: Length of Validation Set: 89 +2016-09-06 16:14:36,159 INFO: Done: Determine folds +2016-09-06 16:14:36,159 INFO: Start: Learning with Fusion and 5 folds +2016-09-06 16:14:36,159 INFO: Start: Randomsearching best settings for monoview classifiers +2016-09-06 16:14:36,160 DEBUG: Start: Random search for DecisionTree with 30 iterations +2016-09-06 16:14:36,162 INFO: Info: Length of Learning Sets: 169 +2016-09-06 16:14:36,162 INFO: Info: Length of Testing Sets: 42 +2016-09-06 16:14:36,162 INFO: Info: Length of Validation Set: 89 +2016-09-06 16:14:36,162 INFO: Done: Determine folds +2016-09-06 16:14:36,162 INFO: Start: Learning with Fusion and 5 folds +2016-09-06 16:14:36,162 INFO: Start: Randomsearching best settings for monoview classifiers +2016-09-06 16:14:36,162 DEBUG: Start: Random search for DecisionTree with 30 iterations +2016-09-06 16:14:36,589 DEBUG: Done: Random search for DecisionTree +2016-09-06 16:14:36,589 DEBUG: Start: Random search for DecisionTree with 30 iterations +2016-09-06 16:14:36,594 DEBUG: Done: Random search for DecisionTree +2016-09-06 16:14:36,594 DEBUG: Start: Random search for DecisionTree with 30 iterations +2016-09-06 16:14:37,006 DEBUG: Done: Random search for DecisionTree +2016-09-06 16:14:37,006 DEBUG: Start: Random search for DecisionTree with 30 iterations +2016-09-06 16:14:37,012 DEBUG: Done: Random search for DecisionTree +2016-09-06 16:14:37,012 DEBUG: Start: Random search for DecisionTree with 30 iterations +2016-09-06 16:14:37,423 DEBUG: Done: Random search for DecisionTree +2016-09-06 16:14:37,423 DEBUG: Start: Random search for DecisionTree with 30 iterations +2016-09-06 16:14:37,429 DEBUG: Done: Random search for DecisionTree +2016-09-06 16:14:37,429 DEBUG: Start: Random search for DecisionTree with 30 iterations diff --git a/Code/MonoMutliViewClassifiers/Results/20160906-161457-CMultiV-Benchmark-Methyl_MiRNA__RNASeq_Clinic-Fake-LOG.log b/Code/MonoMutliViewClassifiers/Results/20160906-161457-CMultiV-Benchmark-Methyl_MiRNA__RNASeq_Clinic-Fake-LOG.log new file mode 100644 index 0000000000000000000000000000000000000000..cd3713da6bdd48e685e1a2a5d11509f7380a4cf2 --- /dev/null +++ b/Code/MonoMutliViewClassifiers/Results/20160906-161457-CMultiV-Benchmark-Methyl_MiRNA__RNASeq_Clinic-Fake-LOG.log @@ -0,0 +1,352 @@ +2016-09-06 16:14:57,297 DEBUG: Start: Creating 2 temporary datasets for multiprocessing +2016-09-06 16:14:57,298 WARNING: WARNING : /!\ This may use a lot of HDD storage space : 0.00015446875 Gbytes /!\ +2016-09-06 16:15:02,311 DEBUG: Start: Creating datasets for multiprocessing +2016-09-06 16:15:02,313 INFO: Start: Finding all available mono- & multiview algorithms +2016-09-06 16:15:02,362 INFO: ### Main Programm for Multiview Classification +2016-09-06 16:15:02,362 INFO: ### Classification - Database : Fake ; Views : Methyl, MiRNA_, RNASeq, Clinic ; Algorithm : Fusion ; Cores : 1 +2016-09-06 16:15:02,362 INFO: ### Main Programm for Multiview Classification +2016-09-06 16:15:02,363 INFO: ### Classification - Database : Fake ; Views : Methyl, MiRNA_, RNASeq, Clinic ; Algorithm : Fusion ; Cores : 1 +2016-09-06 16:15:02,363 INFO: Info: Shape of View0 :(300, 15) +2016-09-06 16:15:02,363 INFO: Info: Shape of View0 :(300, 15) +2016-09-06 16:15:02,363 INFO: Info: Shape of View1 :(300, 16) +2016-09-06 16:15:02,364 INFO: Info: Shape of View1 :(300, 16) +2016-09-06 16:15:02,364 INFO: Info: Shape of View2 :(300, 14) +2016-09-06 16:15:02,364 INFO: Info: Shape of View2 :(300, 14) +2016-09-06 16:15:02,364 INFO: Info: Shape of View3 :(300, 13) +2016-09-06 16:15:02,364 INFO: Done: Read Database Files +2016-09-06 16:15:02,365 INFO: Start: Determine validation split for ratio 0.7 +2016-09-06 16:15:02,365 INFO: Info: Shape of View3 :(300, 13) +2016-09-06 16:15:02,365 INFO: Done: Read Database Files +2016-09-06 16:15:02,365 INFO: Start: Determine validation split for ratio 0.7 +2016-09-06 16:15:02,369 INFO: Done: Determine validation split +2016-09-06 16:15:02,369 INFO: Start: Determine 5 folds +2016-09-06 16:15:02,369 INFO: Done: Determine validation split +2016-09-06 16:15:02,369 INFO: Start: Determine 5 folds +2016-09-06 16:15:02,375 INFO: Info: Length of Learning Sets: 169 +2016-09-06 16:15:02,375 INFO: Info: Length of Testing Sets: 41 +2016-09-06 16:15:02,375 INFO: Info: Length of Validation Set: 90 +2016-09-06 16:15:02,375 INFO: Done: Determine folds +2016-09-06 16:15:02,375 INFO: Start: Learning with Fusion and 5 folds +2016-09-06 16:15:02,375 INFO: Start: Randomsearching best settings for monoview classifiers +2016-09-06 16:15:02,375 DEBUG: Start: Random search for DecisionTree with 30 iterations +2016-09-06 16:15:02,376 INFO: Info: Length of Learning Sets: 169 +2016-09-06 16:15:02,376 INFO: Info: Length of Testing Sets: 41 +2016-09-06 16:15:02,376 INFO: Info: Length of Validation Set: 90 +2016-09-06 16:15:02,376 INFO: Done: Determine folds +2016-09-06 16:15:02,376 INFO: Start: Learning with Fusion and 5 folds +2016-09-06 16:15:02,376 INFO: Start: Randomsearching best settings for monoview classifiers +2016-09-06 16:15:02,377 DEBUG: Start: Random search for DecisionTree with 30 iterations +2016-09-06 16:15:02,805 DEBUG: Done: Random search for DecisionTree +2016-09-06 16:15:02,805 DEBUG: Start: Random search for DecisionTree with 30 iterations +2016-09-06 16:15:02,809 DEBUG: Done: Random search for DecisionTree +2016-09-06 16:15:02,809 DEBUG: Start: Random search for DecisionTree with 30 iterations +2016-09-06 16:15:03,243 DEBUG: Done: Random search for DecisionTree +2016-09-06 16:15:03,243 DEBUG: Start: Random search for DecisionTree with 30 iterations +2016-09-06 16:15:03,311 DEBUG: Done: Random search for DecisionTree +2016-09-06 16:15:03,311 DEBUG: Start: Random search for DecisionTree with 30 iterations +2016-09-06 16:15:03,666 DEBUG: Done: Random search for DecisionTree +2016-09-06 16:15:03,666 DEBUG: Start: Random search for DecisionTree with 30 iterations +2016-09-06 16:15:03,737 DEBUG: Done: Random search for DecisionTree +2016-09-06 16:15:03,737 DEBUG: Start: Random search for DecisionTree with 30 iterations +2016-09-06 16:15:04,086 DEBUG: Done: Random search for DecisionTree +2016-09-06 16:15:04,161 DEBUG: Done: Random search for DecisionTree +2016-09-06 16:15:05,703 INFO: Done: Randomsearching best settings for monoview classifiers +2016-09-06 16:15:05,703 INFO: Start: Classification +2016-09-06 16:15:05,703 INFO: Start: Fold number 1 +2016-09-06 16:15:05,733 INFO: Start: Classification +2016-09-06 16:15:05,759 INFO: Done: Fold number 1 +2016-09-06 16:15:05,759 INFO: Start: Fold number 2 +2016-09-06 16:15:05,790 INFO: Start: Classification +2016-09-06 16:15:05,817 INFO: Done: Fold number 2 +2016-09-06 16:15:05,817 INFO: Start: Fold number 3 +2016-09-06 16:15:05,848 INFO: Start: Classification +2016-09-06 16:15:05,875 INFO: Done: Fold number 3 +2016-09-06 16:15:05,876 INFO: Start: Fold number 4 +2016-09-06 16:15:05,906 INFO: Start: Classification +2016-09-06 16:15:05,933 INFO: Done: Fold number 4 +2016-09-06 16:15:05,933 INFO: Start: Fold number 5 +2016-09-06 16:15:05,964 INFO: Start: Classification +2016-09-06 16:15:05,991 INFO: Done: Fold number 5 +2016-09-06 16:15:05,991 INFO: Done: Classification +2016-09-06 16:15:05,991 INFO: Info: Time for Classification: 3[s] +2016-09-06 16:15:05,992 INFO: Start: Result Analysis for Fusion +2016-09-06 16:15:06,172 INFO: Result for Multiview classification with LateFusion + +Average accuracy : + -On Train : 100.0 + -On Test : 48.2926829268 + -On Validation : 49.3333333333 + +Dataset info : + -Database name : Fake + -Labels : Methyl, MiRNA_, RNASeq, Clinic + -Views : Methyl, MiRNA_, RNASeq, Clinic + -5 folds + +Classification configuration : + -Algorithm used : LateFusion with Bayesian Inference using a weight for each view : 0.4761253742, 0.00270164468431, 0.0406662390422, 0.480506742074 + -With monoview classifiers : + - Decision Tree with max_depth : 7 + - Decision Tree with max_depth : 3 + - Decision Tree with max_depth : 24 + - Decision Tree with max_depth : 4 + +Computation time on 1 cores : + Database extraction time : 0:00:00 + Learn Prediction + Fold 1 0:00:03 0:00:00 + Fold 2 0:00:03 0:00:00 + Fold 3 0:00:03 0:00:00 + Fold 4 0:00:03 0:00:00 + Fold 5 0:00:03 0:00:00 + Total 0:00:17 0:00:00 + So a total classification time of 0:00:03. + + +2016-09-06 16:15:06,172 INFO: Done: Result Analysis +2016-09-06 16:15:06,570 INFO: Done: Randomsearching best settings for monoview classifiers +2016-09-06 16:15:06,570 INFO: Start: Classification +2016-09-06 16:15:06,570 INFO: Start: Fold number 1 +2016-09-06 16:15:06,599 INFO: Start: Classification +2016-09-06 16:15:06,667 INFO: Done: Fold number 1 +2016-09-06 16:15:06,668 INFO: Start: Fold number 2 +2016-09-06 16:15:06,697 INFO: Start: Classification +2016-09-06 16:15:06,765 INFO: Done: Fold number 2 +2016-09-06 16:15:06,765 INFO: Start: Fold number 3 +2016-09-06 16:15:06,794 INFO: Start: Classification +2016-09-06 16:15:06,862 INFO: Done: Fold number 3 +2016-09-06 16:15:06,862 INFO: Start: Fold number 4 +2016-09-06 16:15:06,896 INFO: Start: Classification +2016-09-06 16:15:06,964 INFO: Done: Fold number 4 +2016-09-06 16:15:06,964 INFO: Start: Fold number 5 +2016-09-06 16:15:06,993 INFO: Start: Classification +2016-09-06 16:15:07,061 INFO: Done: Fold number 5 +2016-09-06 16:15:07,061 INFO: Done: Classification +2016-09-06 16:15:07,061 INFO: Info: Time for Classification: 4[s] +2016-09-06 16:15:07,061 INFO: Start: Result Analysis for Fusion +2016-09-06 16:15:07,194 INFO: Result for Multiview classification with LateFusion + +Average accuracy : + -On Train : 100.0 + -On Test : 46.8292682927 + -On Validation : 44.0 + +Dataset info : + -Database name : Fake + -Labels : Methyl, MiRNA_, RNASeq, Clinic + -Views : Methyl, MiRNA_, RNASeq, Clinic + -5 folds + +Classification configuration : + -Algorithm used : LateFusion with Majority Voting + -With monoview classifiers : + - Decision Tree with max_depth : 7 + - Decision Tree with max_depth : 3 + - Decision Tree with max_depth : 24 + - Decision Tree with max_depth : 4 + +Computation time on 1 cores : + Database extraction time : 0:00:00 + Learn Prediction + Fold 1 0:00:04 0:00:00 + Fold 2 0:00:04 0:00:00 + Fold 3 0:00:04 0:00:00 + Fold 4 0:00:04 0:00:00 + Fold 5 0:00:04 0:00:00 + Total 0:00:22 0:00:00 + So a total classification time of 0:00:04. + + +2016-09-06 16:15:07,195 INFO: Done: Result Analysis +2016-09-06 16:15:07,318 INFO: ### Main Programm for Multiview Classification +2016-09-06 16:15:07,318 INFO: ### Main Programm for Multiview Classification +2016-09-06 16:15:07,319 INFO: ### Classification - Database : Fake ; Views : Methyl, MiRNA_, RNASeq, Clinic ; Algorithm : Fusion ; Cores : 1 +2016-09-06 16:15:07,319 INFO: ### Classification - Database : Fake ; Views : Methyl, MiRNA_, RNASeq, Clinic ; Algorithm : Fusion ; Cores : 1 +2016-09-06 16:15:07,319 INFO: Info: Shape of View0 :(300, 15) +2016-09-06 16:15:07,319 INFO: Info: Shape of View0 :(300, 15) +2016-09-06 16:15:07,320 INFO: Info: Shape of View1 :(300, 16) +2016-09-06 16:15:07,320 INFO: Info: Shape of View1 :(300, 16) +2016-09-06 16:15:07,321 INFO: Info: Shape of View2 :(300, 14) +2016-09-06 16:15:07,321 INFO: Info: Shape of View2 :(300, 14) +2016-09-06 16:15:07,321 INFO: Info: Shape of View3 :(300, 13) +2016-09-06 16:15:07,321 INFO: Info: Shape of View3 :(300, 13) +2016-09-06 16:15:07,321 INFO: Done: Read Database Files +2016-09-06 16:15:07,321 INFO: Done: Read Database Files +2016-09-06 16:15:07,322 INFO: Start: Determine validation split for ratio 0.7 +2016-09-06 16:15:07,322 INFO: Start: Determine validation split for ratio 0.7 +2016-09-06 16:15:07,327 INFO: Done: Determine validation split +2016-09-06 16:15:07,328 INFO: Done: Determine validation split +2016-09-06 16:15:07,328 INFO: Start: Determine 5 folds +2016-09-06 16:15:07,328 INFO: Start: Determine 5 folds +2016-09-06 16:15:07,339 INFO: Info: Length of Learning Sets: 169 +2016-09-06 16:15:07,339 INFO: Info: Length of Testing Sets: 41 +2016-09-06 16:15:07,339 INFO: Info: Length of Validation Set: 90 +2016-09-06 16:15:07,339 INFO: Done: Determine folds +2016-09-06 16:15:07,339 INFO: Start: Learning with Fusion and 5 folds +2016-09-06 16:15:07,339 INFO: Start: Randomsearching best settings for monoview classifiers +2016-09-06 16:15:07,339 DEBUG: Start: Random search for DecisionTree with 30 iterations +2016-09-06 16:15:07,344 INFO: Info: Length of Learning Sets: 169 +2016-09-06 16:15:07,345 INFO: Info: Length of Testing Sets: 41 +2016-09-06 16:15:07,345 INFO: Info: Length of Validation Set: 90 +2016-09-06 16:15:07,345 INFO: Done: Determine folds +2016-09-06 16:15:07,345 INFO: Start: Learning with Fusion and 5 folds +2016-09-06 16:15:07,345 INFO: Start: Randomsearching best settings for monoview classifiers +2016-09-06 16:15:07,345 DEBUG: Start: Random search for DecisionTree with 30 iterations +2016-09-06 16:15:08,051 DEBUG: Done: Random search for DecisionTree +2016-09-06 16:15:08,051 DEBUG: Start: Random search for DecisionTree with 30 iterations +2016-09-06 16:15:08,056 DEBUG: Done: Random search for DecisionTree +2016-09-06 16:15:08,057 DEBUG: Start: Random search for DecisionTree with 30 iterations +2016-09-06 16:15:08,762 DEBUG: Done: Random search for DecisionTree +2016-09-06 16:15:08,762 DEBUG: Start: Random search for DecisionTree with 30 iterations +2016-09-06 16:15:08,764 DEBUG: Done: Random search for DecisionTree +2016-09-06 16:15:08,765 DEBUG: Start: Random search for DecisionTree with 30 iterations +2016-09-06 16:15:09,471 DEBUG: Done: Random search for DecisionTree +2016-09-06 16:15:09,471 DEBUG: Start: Random search for DecisionTree with 30 iterations +2016-09-06 16:15:09,474 DEBUG: Done: Random search for DecisionTree +2016-09-06 16:15:09,474 DEBUG: Start: Random search for DecisionTree with 30 iterations +2016-09-06 16:15:10,189 DEBUG: Done: Random search for DecisionTree +2016-09-06 16:15:10,190 INFO: Done: Randomsearching best settings for monoview classifiers +2016-09-06 16:15:10,190 INFO: Start: Classification +2016-09-06 16:15:10,190 INFO: Start: Fold number 1 +2016-09-06 16:15:10,192 DEBUG: Done: Random search for DecisionTree +2016-09-06 16:15:10,252 INFO: Start: Classification +2016-09-06 16:15:10,280 INFO: Done: Fold number 1 +2016-09-06 16:15:10,280 INFO: Start: Fold number 2 +2016-09-06 16:15:10,327 INFO: Start: Classification +2016-09-06 16:15:10,355 INFO: Done: Fold number 2 +2016-09-06 16:15:10,355 INFO: Start: Fold number 3 +2016-09-06 16:15:10,402 INFO: Start: Classification +2016-09-06 16:15:10,430 INFO: Done: Fold number 3 +2016-09-06 16:15:10,430 INFO: Start: Fold number 4 +2016-09-06 16:15:10,477 INFO: Start: Classification +2016-09-06 16:15:10,505 INFO: Done: Fold number 4 +2016-09-06 16:15:10,505 INFO: Start: Fold number 5 +2016-09-06 16:15:10,553 INFO: Start: Classification +2016-09-06 16:15:10,581 INFO: Done: Fold number 5 +2016-09-06 16:15:10,581 INFO: Done: Classification +2016-09-06 16:15:10,581 INFO: Info: Time for Classification: 3[s] +2016-09-06 16:15:10,581 INFO: Start: Result Analysis for Fusion +2016-09-06 16:15:10,751 INFO: Result for Multiview classification with LateFusion + +Average accuracy : + -On Train : 100.0 + -On Test : 44.8780487805 + -On Validation : 50.6666666667 + +Dataset info : + -Database name : Fake + -Labels : Methyl, MiRNA_, RNASeq, Clinic + -Views : Methyl, MiRNA_, RNASeq, Clinic + -5 folds + +Classification configuration : + -Algorithm used : LateFusion with SVM for linear + -With monoview classifiers : + - Decision Tree with max_depth : 7 + - Decision Tree with max_depth : 3 + - Decision Tree with max_depth : 24 + - Decision Tree with max_depth : 4 + +Computation time on 1 cores : + Database extraction time : 0:00:00 + Learn Prediction + Fold 1 0:00:02 0:00:00 + Fold 2 0:00:02 0:00:00 + Fold 3 0:00:03 0:00:00 + Fold 4 0:00:03 0:00:00 + Fold 5 0:00:03 0:00:00 + Total 0:00:15 0:00:00 + So a total classification time of 0:00:03. + + +2016-09-06 16:15:10,765 INFO: Done: Result Analysis +2016-09-06 16:15:11,829 INFO: Done: Randomsearching best settings for monoview classifiers +2016-09-06 16:15:11,830 INFO: Start: Classification +2016-09-06 16:15:11,830 INFO: Start: Fold number 1 +2016-09-06 16:15:11,858 INFO: Start: Classification +2016-09-06 16:15:11,888 INFO: Done: Fold number 1 +2016-09-06 16:15:11,888 INFO: Start: Fold number 2 +2016-09-06 16:15:11,917 INFO: Start: Classification +2016-09-06 16:15:11,947 INFO: Done: Fold number 2 +2016-09-06 16:15:11,947 INFO: Start: Fold number 3 +2016-09-06 16:15:11,975 INFO: Start: Classification +2016-09-06 16:15:12,005 INFO: Done: Fold number 3 +2016-09-06 16:15:12,005 INFO: Start: Fold number 4 +2016-09-06 16:15:12,033 INFO: Start: Classification +2016-09-06 16:15:12,063 INFO: Done: Fold number 4 +2016-09-06 16:15:12,064 INFO: Start: Fold number 5 +2016-09-06 16:15:12,092 INFO: Start: Classification +2016-09-06 16:15:12,122 INFO: Done: Fold number 5 +2016-09-06 16:15:12,122 INFO: Done: Classification +2016-09-06 16:15:12,122 INFO: Info: Time for Classification: 4[s] +2016-09-06 16:15:12,122 INFO: Start: Result Analysis for Fusion +2016-09-06 16:15:12,264 INFO: Result for Multiview classification with LateFusion + +Average accuracy : + -On Train : 53.2544378698 + -On Test : 53.6585365854 + -On Validation : 53.3333333333 + +Dataset info : + -Database name : Fake + -Labels : Methyl, MiRNA_, RNASeq, Clinic + -Views : Methyl, MiRNA_, RNASeq, Clinic + -5 folds + +Classification configuration : + -Algorithm used : LateFusion with Weighted linear using a weight for each view : 1.0, 0.111861495999, 0.371763764635, 0.0125470094031 + -With monoview classifiers : + - Decision Tree with max_depth : 7 + - Decision Tree with max_depth : 3 + - Decision Tree with max_depth : 24 + - Decision Tree with max_depth : 4 + +Computation time on 1 cores : + Database extraction time : 0:00:00 + Learn Prediction + Fold 1 0:00:04 0:00:00 + Fold 2 0:00:04 0:00:00 + Fold 3 0:00:04 0:00:00 + Fold 4 0:00:04 0:00:00 + Fold 5 0:00:04 0:00:00 + Total 0:00:23 0:00:00 + So a total classification time of 0:00:04. + + +2016-09-06 16:15:12,264 INFO: Done: Result Analysis +2016-09-06 16:15:12,376 INFO: ### Main Programm for Multiview Classification +2016-09-06 16:15:12,376 INFO: ### Classification - Database : Fake ; Views : Methyl, MiRNA_, RNASeq, Clinic ; Algorithm : Fusion ; Cores : 1 +2016-09-06 16:15:12,377 INFO: ### Main Programm for Multiview Classification +2016-09-06 16:15:12,377 INFO: ### Classification - Database : Fake ; Views : Methyl, MiRNA_, RNASeq, Clinic ; Algorithm : Fusion ; Cores : 1 +2016-09-06 16:15:12,377 INFO: Info: Shape of View0 :(300, 15) +2016-09-06 16:15:12,378 INFO: Info: Shape of View0 :(300, 15) +2016-09-06 16:15:12,378 INFO: Info: Shape of View1 :(300, 16) +2016-09-06 16:15:12,379 INFO: Info: Shape of View1 :(300, 16) +2016-09-06 16:15:12,379 INFO: Info: Shape of View2 :(300, 14) +2016-09-06 16:15:12,380 INFO: Info: Shape of View2 :(300, 14) +2016-09-06 16:15:12,380 INFO: Info: Shape of View3 :(300, 13) +2016-09-06 16:15:12,380 INFO: Done: Read Database Files +2016-09-06 16:15:12,380 INFO: Start: Determine validation split for ratio 0.7 +2016-09-06 16:15:12,381 INFO: Info: Shape of View3 :(300, 13) +2016-09-06 16:15:12,381 INFO: Done: Read Database Files +2016-09-06 16:15:12,381 INFO: Start: Determine validation split for ratio 0.7 +2016-09-06 16:15:12,385 INFO: Done: Determine validation split +2016-09-06 16:15:12,385 INFO: Start: Determine 5 folds +2016-09-06 16:15:12,385 INFO: Done: Determine validation split +2016-09-06 16:15:12,385 INFO: Start: Determine 5 folds +2016-09-06 16:15:12,392 INFO: Info: Length of Learning Sets: 169 +2016-09-06 16:15:12,392 INFO: Info: Length of Testing Sets: 41 +2016-09-06 16:15:12,392 INFO: Info: Length of Validation Set: 90 +2016-09-06 16:15:12,392 INFO: Done: Determine folds +2016-09-06 16:15:12,392 INFO: Start: Learning with Fusion and 5 folds +2016-09-06 16:15:12,393 INFO: Start: Randomsearching best settings for monoview classifiers +2016-09-06 16:15:12,393 DEBUG: Start: Random search for DecisionTree with 30 iterations +2016-09-06 16:15:12,393 INFO: Info: Length of Learning Sets: 169 +2016-09-06 16:15:12,393 INFO: Info: Length of Testing Sets: 41 +2016-09-06 16:15:12,393 INFO: Info: Length of Validation Set: 90 +2016-09-06 16:15:12,393 INFO: Done: Determine folds +2016-09-06 16:15:12,393 INFO: Start: Learning with Fusion and 5 folds +2016-09-06 16:15:12,393 INFO: Start: Randomsearching best settings for monoview classifiers +2016-09-06 16:15:12,393 DEBUG: Start: Random search for Adaboost with 30 iterations +2016-09-06 16:15:12,822 DEBUG: Done: Random search for DecisionTree +2016-09-06 16:15:15,451 DEBUG: Done: Random search for Adaboost diff --git a/Code/MonoMutliViewClassifiers/Results/20160906-161506Results-Fusion-LateFusion-BayesianInference-DecisionTree-DecisionTree-DecisionTree-DecisionTree-Methyl-MiRNA_-RNASeq-Clinic-MiRNA_-RNASeq-Clinic-Methyl-learnRate0.7-Fake.txt b/Code/MonoMutliViewClassifiers/Results/20160906-161506Results-Fusion-LateFusion-BayesianInference-DecisionTree-DecisionTree-DecisionTree-DecisionTree-Methyl-MiRNA_-RNASeq-Clinic-MiRNA_-RNASeq-Clinic-Methyl-learnRate0.7-Fake.txt new file mode 100644 index 0000000000000000000000000000000000000000..05f6d2404a2dcd983b37b1df7c9d17ab53b37cbf --- /dev/null +++ b/Code/MonoMutliViewClassifiers/Results/20160906-161506Results-Fusion-LateFusion-BayesianInference-DecisionTree-DecisionTree-DecisionTree-DecisionTree-Methyl-MiRNA_-RNASeq-Clinic-MiRNA_-RNASeq-Clinic-Methyl-learnRate0.7-Fake.txt @@ -0,0 +1,32 @@ + Result for Multiview classification with LateFusion + +Average accuracy : + -On Train : 100.0 + -On Test : 48.2926829268 + -On Validation : 49.3333333333 + +Dataset info : + -Database name : Fake + -Labels : Methyl, MiRNA_, RNASeq, Clinic + -Views : Methyl, MiRNA_, RNASeq, Clinic + -5 folds + +Classification configuration : + -Algorithm used : LateFusion with Bayesian Inference using a weight for each view : 0.4761253742, 0.00270164468431, 0.0406662390422, 0.480506742074 + -With monoview classifiers : + - Decision Tree with max_depth : 7 + - Decision Tree with max_depth : 3 + - Decision Tree with max_depth : 24 + - Decision Tree with max_depth : 4 + +Computation time on 1 cores : + Database extraction time : 0:00:00 + Learn Prediction + Fold 1 0:00:03 0:00:00 + Fold 2 0:00:03 0:00:00 + Fold 3 0:00:03 0:00:00 + Fold 4 0:00:03 0:00:00 + Fold 5 0:00:03 0:00:00 + Total 0:00:17 0:00:00 + So a total classification time of 0:00:03. + diff --git a/Code/MonoMutliViewClassifiers/Results/20160906-161507Results-Fusion-LateFusion-MajorityVoting-DecisionTree-DecisionTree-DecisionTree-DecisionTree-Methyl-MiRNA_-RNASeq-Clinic-MiRNA_-RNASeq-Clinic-Methyl-learnRate0.7-Fake.txt b/Code/MonoMutliViewClassifiers/Results/20160906-161507Results-Fusion-LateFusion-MajorityVoting-DecisionTree-DecisionTree-DecisionTree-DecisionTree-Methyl-MiRNA_-RNASeq-Clinic-MiRNA_-RNASeq-Clinic-Methyl-learnRate0.7-Fake.txt new file mode 100644 index 0000000000000000000000000000000000000000..83b87a9205a913e9cdd6483808f8d96f2c702d32 --- /dev/null +++ b/Code/MonoMutliViewClassifiers/Results/20160906-161507Results-Fusion-LateFusion-MajorityVoting-DecisionTree-DecisionTree-DecisionTree-DecisionTree-Methyl-MiRNA_-RNASeq-Clinic-MiRNA_-RNASeq-Clinic-Methyl-learnRate0.7-Fake.txt @@ -0,0 +1,32 @@ + Result for Multiview classification with LateFusion + +Average accuracy : + -On Train : 100.0 + -On Test : 46.8292682927 + -On Validation : 44.0 + +Dataset info : + -Database name : Fake + -Labels : Methyl, MiRNA_, RNASeq, Clinic + -Views : Methyl, MiRNA_, RNASeq, Clinic + -5 folds + +Classification configuration : + -Algorithm used : LateFusion with Majority Voting + -With monoview classifiers : + - Decision Tree with max_depth : 7 + - Decision Tree with max_depth : 3 + - Decision Tree with max_depth : 24 + - Decision Tree with max_depth : 4 + +Computation time on 1 cores : + Database extraction time : 0:00:00 + Learn Prediction + Fold 1 0:00:04 0:00:00 + Fold 2 0:00:04 0:00:00 + Fold 3 0:00:04 0:00:00 + Fold 4 0:00:04 0:00:00 + Fold 5 0:00:04 0:00:00 + Total 0:00:22 0:00:00 + So a total classification time of 0:00:04. + diff --git a/Code/MonoMutliViewClassifiers/Results/20160906-161510Results-Fusion-LateFusion-SVMForLinear-DecisionTree-DecisionTree-DecisionTree-DecisionTree-Methyl-MiRNA_-RNASeq-Clinic-MiRNA_-RNASeq-Clinic-Methyl-learnRate0.7-Fake.txt b/Code/MonoMutliViewClassifiers/Results/20160906-161510Results-Fusion-LateFusion-SVMForLinear-DecisionTree-DecisionTree-DecisionTree-DecisionTree-Methyl-MiRNA_-RNASeq-Clinic-MiRNA_-RNASeq-Clinic-Methyl-learnRate0.7-Fake.txt new file mode 100644 index 0000000000000000000000000000000000000000..c190e8d144c1431ffbef8cd728c7e9311e78d508 --- /dev/null +++ b/Code/MonoMutliViewClassifiers/Results/20160906-161510Results-Fusion-LateFusion-SVMForLinear-DecisionTree-DecisionTree-DecisionTree-DecisionTree-Methyl-MiRNA_-RNASeq-Clinic-MiRNA_-RNASeq-Clinic-Methyl-learnRate0.7-Fake.txt @@ -0,0 +1,32 @@ + Result for Multiview classification with LateFusion + +Average accuracy : + -On Train : 100.0 + -On Test : 44.8780487805 + -On Validation : 50.6666666667 + +Dataset info : + -Database name : Fake + -Labels : Methyl, MiRNA_, RNASeq, Clinic + -Views : Methyl, MiRNA_, RNASeq, Clinic + -5 folds + +Classification configuration : + -Algorithm used : LateFusion with SVM for linear + -With monoview classifiers : + - Decision Tree with max_depth : 7 + - Decision Tree with max_depth : 3 + - Decision Tree with max_depth : 24 + - Decision Tree with max_depth : 4 + +Computation time on 1 cores : + Database extraction time : 0:00:00 + Learn Prediction + Fold 1 0:00:02 0:00:00 + Fold 2 0:00:02 0:00:00 + Fold 3 0:00:03 0:00:00 + Fold 4 0:00:03 0:00:00 + Fold 5 0:00:03 0:00:00 + Total 0:00:15 0:00:00 + So a total classification time of 0:00:03. + diff --git a/Code/MonoMutliViewClassifiers/Results/20160906-161512Results-Fusion-LateFusion-WeightedLinear-DecisionTree-DecisionTree-DecisionTree-DecisionTree-Methyl-MiRNA_-RNASeq-Clinic-MiRNA_-RNASeq-Clinic-Methyl-learnRate0.7-Fake.txt b/Code/MonoMutliViewClassifiers/Results/20160906-161512Results-Fusion-LateFusion-WeightedLinear-DecisionTree-DecisionTree-DecisionTree-DecisionTree-Methyl-MiRNA_-RNASeq-Clinic-MiRNA_-RNASeq-Clinic-Methyl-learnRate0.7-Fake.txt new file mode 100644 index 0000000000000000000000000000000000000000..5e8e31ae0495caadf8b277decbca8b06dca480e3 --- /dev/null +++ b/Code/MonoMutliViewClassifiers/Results/20160906-161512Results-Fusion-LateFusion-WeightedLinear-DecisionTree-DecisionTree-DecisionTree-DecisionTree-Methyl-MiRNA_-RNASeq-Clinic-MiRNA_-RNASeq-Clinic-Methyl-learnRate0.7-Fake.txt @@ -0,0 +1,32 @@ + Result for Multiview classification with LateFusion + +Average accuracy : + -On Train : 53.2544378698 + -On Test : 53.6585365854 + -On Validation : 53.3333333333 + +Dataset info : + -Database name : Fake + -Labels : Methyl, MiRNA_, RNASeq, Clinic + -Views : Methyl, MiRNA_, RNASeq, Clinic + -5 folds + +Classification configuration : + -Algorithm used : LateFusion with Weighted linear using a weight for each view : 1.0, 0.111861495999, 0.371763764635, 0.0125470094031 + -With monoview classifiers : + - Decision Tree with max_depth : 7 + - Decision Tree with max_depth : 3 + - Decision Tree with max_depth : 24 + - Decision Tree with max_depth : 4 + +Computation time on 1 cores : + Database extraction time : 0:00:00 + Learn Prediction + Fold 1 0:00:04 0:00:00 + Fold 2 0:00:04 0:00:00 + Fold 3 0:00:04 0:00:00 + Fold 4 0:00:04 0:00:00 + Fold 5 0:00:04 0:00:00 + Total 0:00:23 0:00:00 + So a total classification time of 0:00:04. +