From 7fb057942628d853f1f11c9d81a7513c91710d9a Mon Sep 17 00:00:00 2001
From: Baptiste Bauvin <baptiste.bauvin.1@ulaval.ca>
Date: Thu, 19 Oct 2017 15:10:30 -0400
Subject: [PATCH] Added interpret for adaboost and DT
---
.../Monoview/ExecClassifMonoView.py | 2 +-
.../Monoview/analyzeResult.py | 12 +++++---
.../MonoviewClassifiers/Adaboost.py | 18 ++++++++++++
.../MonoviewClassifiers/DecisionTree.py | 28 +++++++++++++++++--
.../Methods/LateFusionPackage/SCMForLinear.py | 11 --------
Code/Versions.py | 7 +++++
6 files changed, 59 insertions(+), 19 deletions(-)
diff --git a/Code/MonoMultiViewClassifiers/Monoview/ExecClassifMonoView.py b/Code/MonoMultiViewClassifiers/Monoview/ExecClassifMonoView.py
index 312f4fe4..e8d0b308 100644
--- a/Code/MonoMultiViewClassifiers/Monoview/ExecClassifMonoView.py
+++ b/Code/MonoMultiViewClassifiers/Monoview/ExecClassifMonoView.py
@@ -113,7 +113,7 @@ def ExecMonoview(directory, X, Y, name, labelsNames, classificationIndices, KFol
hyperParamSearch, metrics, nIter, feat, CL_type,
clKWARGS, labelsNames, X.shape,
y_train, y_train_pred, y_test, y_test_pred, t_end,
- randomState)
+ randomState, cl_res, outputFileName)
cl_desc = [value for key, value in sorted(clKWARGS.iteritems())]
logging.debug("Done:\t Getting Results")
logging.info(stringAnalysis)
diff --git a/Code/MonoMultiViewClassifiers/Monoview/analyzeResult.py b/Code/MonoMultiViewClassifiers/Monoview/analyzeResult.py
index 998c3c86..e86c88d3 100644
--- a/Code/MonoMultiViewClassifiers/Monoview/analyzeResult.py
+++ b/Code/MonoMultiViewClassifiers/Monoview/analyzeResult.py
@@ -16,7 +16,7 @@ def getDBConfigString(name, feat, classificationIndices, shape, classLabelsNames
return dbConfigString
-def getClassifierConfigString(CL_type, gridSearch, nbCores, nIter, clKWARGS):
+def getClassifierConfigString(CL_type, gridSearch, nbCores, nIter, clKWARGS, classififer, directory):
classifierModule = getattr(MonoviewClassifiers, CL_type)
classifierConfigString = "Classifier configuration : \n"
classifierConfigString += "\t- " + classifierModule.getConfig(clKWARGS)[5:] + "\n"
@@ -24,7 +24,8 @@ def getClassifierConfigString(CL_type, gridSearch, nbCores, nIter, clKWARGS):
if gridSearch:
classifierConfigString += "\t- Got configuration using randomized search with " + str(nIter) + " iterations \n"
classifierConfigString += "\n\n"
- return classifierConfigString
+ classifierInterpretString = classifierModule.getInterpret(classififer, directory)
+ return classifierConfigString, classifierInterpretString
def getMetricScore(metric, y_train, y_train_pred, y_test, y_test_pred):
@@ -43,7 +44,7 @@ def getMetricScore(metric, y_train, y_train_pred, y_test, y_test_pred):
def execute(name, learningRate, KFolds, nbCores, gridSearch, metrics, nIter, feat, CL_type, clKWARGS, classLabelsNames,
- shape, y_train, y_train_pred, y_test, y_test_pred, time, randomState):
+ shape, y_train, y_train_pred, y_test, y_test_pred, time, randomState, classifier, directory):
metricsScores = {}
metricModule = getattr(Metrics, metrics[0][0])
trainScore = metricModule.score(y_train, y_train_pred)
@@ -53,7 +54,8 @@ def execute(name, learningRate, KFolds, nbCores, gridSearch, metrics, nIter, fea
stringAnalysis += metrics[0][0] + " on train : " + str(trainScore) + "\n" + metrics[0][0] + " on test : " + str(
testScore) + "\n\n"
stringAnalysis += getDBConfigString(name, feat, learningRate, shape, classLabelsNames, KFolds)
- stringAnalysis += getClassifierConfigString(CL_type, gridSearch, nbCores, nIter, clKWARGS)
+ classifierConfigString, classifierIntepretString = getClassifierConfigString(CL_type, gridSearch, nbCores, nIter, clKWARGS, classifier, directory)
+ stringAnalysis += classifierConfigString
for metric in metrics:
stringAnalysis += getMetricScore(metric, y_train, y_train_pred, y_test, y_test_pred)
if metric[1] is not None:
@@ -63,6 +65,8 @@ def execute(name, learningRate, KFolds, nbCores, gridSearch, metrics, nIter, fea
metricsScores[metric[0]] = [getattr(Metrics, metric[0]).score(y_train, y_train_pred),
getattr(Metrics, metric[0]).score(y_test, y_test_pred)]
stringAnalysis += "\n\n Classification took " + str(hms(seconds=int(time)))
+ stringAnalysis += "\n\n Classifier Interpretation : \n"
+ stringAnalysis+= classifierIntepretString
imageAnalysis = {}
return stringAnalysis, imageAnalysis, metricsScores
diff --git a/Code/MonoMultiViewClassifiers/MonoviewClassifiers/Adaboost.py b/Code/MonoMultiViewClassifiers/MonoviewClassifiers/Adaboost.py
index cad2b46b..66a39edd 100644
--- a/Code/MonoMultiViewClassifiers/MonoviewClassifiers/Adaboost.py
+++ b/Code/MonoMultiViewClassifiers/MonoviewClassifiers/Adaboost.py
@@ -4,6 +4,7 @@ from sklearn.model_selection import RandomizedSearchCV
from sklearn.tree import DecisionTreeClassifier
from scipy.stats import randint
import numpy as np
+import cPickle
from .. import Metrics
from ..utils.HyperParameterSearch import genHeatMaps
@@ -79,3 +80,20 @@ def getConfig(config):
except:
return "\n\t\t- Adaboost with num_esimators : " + str(config["0"]) + ", base_estimators : " + str(
config["1"])
+
+def getInterpret(classifier, directory):
+ featureImportances = classifier.feature_importances_
+ sortedArgs = np.argsort(-featureImportances)
+ featureImportancesSorted = featureImportances[sortedArgs][:50]
+ featureIndicesSorted = sortedArgs[:50]
+ featuresImportancesDict = dict((featureIndex, featureImportance)
+ for featureIndex, featureImportance in enumerate(featureImportances)
+ if featureImportance != 0)
+ with open(directory+'-feature_importances.pickle', 'wb') as handle:
+ cPickle.dump(featuresImportancesDict, handle)
+ interpretString = "Feature importances : \n"
+ for featureIndex, featureImportance in zip(featureIndicesSorted, featureImportancesSorted):
+ if featureImportance>0:
+ interpretString+="- Feature index : "+str(featureIndex)+\
+ ", feature importance : "+str(featureImportance)+"\n"
+ return interpretString
\ No newline at end of file
diff --git a/Code/MonoMultiViewClassifiers/MonoviewClassifiers/DecisionTree.py b/Code/MonoMultiViewClassifiers/MonoviewClassifiers/DecisionTree.py
index 9b813c53..152e9e7d 100644
--- a/Code/MonoMultiViewClassifiers/MonoviewClassifiers/DecisionTree.py
+++ b/Code/MonoMultiViewClassifiers/MonoviewClassifiers/DecisionTree.py
@@ -1,8 +1,10 @@
-from sklearn.tree import DecisionTreeClassifier
+from sklearn import tree
from sklearn.pipeline import Pipeline # Pipelining in classification
from sklearn.model_selection import RandomizedSearchCV
from scipy.stats import randint
import numpy as np
+import graphviz
+import cPickle
from .. import Metrics
from ..utils.HyperParameterSearch import genHeatMaps
@@ -20,7 +22,7 @@ def fit(DATASET, CLASS_LABELS, randomState, NB_CORES=1, **kwargs):
maxDepth = int(kwargs['0'])
criterion = kwargs['1']
splitter = kwargs['2']
- classifier = DecisionTreeClassifier(max_depth=maxDepth, criterion=criterion, splitter=splitter,
+ classifier = tree.DecisionTreeClassifier(max_depth=maxDepth, criterion=criterion, splitter=splitter,
random_state=randomState)
classifier.fit(DATASET, CLASS_LABELS)
return classifier
@@ -48,7 +50,7 @@ def getKWARGS(kwargsList):
def randomizedSearch(X_train, y_train, randomState, outputFileName, KFolds=4, nbCores=1,
metric=["accuracy_score", None], nIter=30):
- pipeline_DT = Pipeline([('classifier', DecisionTreeClassifier())])
+ pipeline_DT = Pipeline([('classifier', tree.DecisionTreeClassifier())])
param_DT = {"classifier__max_depth": randint(1, 300),
"classifier__criterion": ["gini", "entropy"],
"classifier__splitter": ["best", "random"]}
@@ -85,3 +87,23 @@ def getConfig(config):
except:
return "\n\t\t- Decision Tree with max_depth : " + str(config["0"]) + ", criterion : " + config[
"1"] + ", splitter : " + config["2"]
+
+def getInterpret(classifier, directory):
+ dot_data = tree.export_graphviz(classifier, out_file=None)
+ graph = graphviz.Source(dot_data)
+ graph.render(directory+"-tree.pdf")
+ featureImportances = classifier.feature_importances_
+ sortedArgs = np.argsort(-featureImportances)
+ featureImportancesSorted = featureImportances[sortedArgs][:50]
+ featureIndicesSorted = sortedArgs[:50]
+ featuresImportancesDict = dict((featureIndex, featureImportance)
+ for featureIndex, featureImportance in enumerate(featureImportances)
+ if featureImportance != 0)
+ with open(directory + '-feature_importances.pickle', 'wb') as handle:
+ cPickle.dump(featuresImportancesDict, handle)
+ interpretString = "Feature importances : \n"
+ for featureIndex, featureImportance in zip(featureIndicesSorted, featureImportancesSorted):
+ if featureImportance > 0:
+ interpretString += "- Feature index : " + str(featureIndex) + \
+ ", feature importance : " + str(featureImportance) + "\n"
+ return interpretString
\ No newline at end of file
diff --git a/Code/MonoMultiViewClassifiers/Multiview/Fusion/Methods/LateFusionPackage/SCMForLinear.py b/Code/MonoMultiViewClassifiers/Multiview/Fusion/Methods/LateFusionPackage/SCMForLinear.py
index 29a879d4..43a22a2f 100644
--- a/Code/MonoMultiViewClassifiers/Multiview/Fusion/Methods/LateFusionPackage/SCMForLinear.py
+++ b/Code/MonoMultiViewClassifiers/Multiview/Fusion/Methods/LateFusionPackage/SCMForLinear.py
@@ -1,22 +1,11 @@
import numpy as np
-import pyscm
-# from pyscm.utils import _pack_binary_bytes_to_ints
-import os
-import h5py
-# from pyscm.binary_attributes.classifications.popcount import inplace_popcount_32, inplace_popcount_64
-# from pyscm.utils import _unpack_binary_bytes_from_ints
from pyscm.scm import SetCoveringMachineClassifier as scm
from sklearn.base import BaseEstimator, ClassifierMixin
-from sklearn.pipeline import Pipeline
-from sklearn.model_selection import RandomizedSearchCV
from sklearn.externals.six import iteritems, iterkeys, itervalues
-from math import ceil
-import random
from sklearn.metrics import accuracy_score
import itertools
-import pkgutil
from ..LateFusion import LateFusionClassifier, getClassifiers, getConfig
from ..... import MonoviewClassifiers
diff --git a/Code/Versions.py b/Code/Versions.py
index 8a83f563..b3a02c11 100644
--- a/Code/Versions.py
+++ b/Code/Versions.py
@@ -96,8 +96,15 @@ def testVersions():
isUpToDate = False
toInstall.append("h5py")
+ try:
+ import graphviz #
+ except:
+ isUpToDate = False
+ toInstall.append("graphviz")
+
if not isUpToDate:
print "You can't run at the moment, please install the following modules : \n"+ "\n".join(toInstall)
+ quit()
if __name__== "__main__":
testVersions()
\ No newline at end of file
--
GitLab