From cf34b931a62d9497bf1ce75476432aa53467eb80 Mon Sep 17 00:00:00 2001
From: Baptiste Bauvin <baptiste.bauvin@lis-lab.fr>
Date: Tue, 21 Jan 2020 14:25:35 +0100
Subject: [PATCH] Corrected cross val multiview
---
config_files/config_test.yml | 18 +++++++++---------
.../weighted_linear_early_fusion.py | 3 ++-
.../result_analysis.py | 15 ++++++++++-----
.../utils/hyper_parameter_search.py | 6 +++---
4 files changed, 24 insertions(+), 18 deletions(-)
diff --git a/config_files/config_test.yml b/config_files/config_test.yml
index f54013c7..d3142c90 100644
--- a/config_files/config_test.yml
+++ b/config_files/config_test.yml
@@ -1,11 +1,11 @@
# The base configuration of the benchmark
Base :
log: True
- name: ["outliers_dset"]
+ name: ["awa-tiger-wolf-all"]
label: "_"
type: ".hdf5"
views:
- pathf: "/home/baptiste/Documents/Datasets/Generated/outliers_dset/"
+ pathf: "/home/baptiste/Documents/Datasets/AWA/base/"
nice: 0
random_state: 42
nb_cores: 1
@@ -18,16 +18,16 @@ Base :
# All the classification-realted configuration options
Classification:
multiclass_method: "oneVersusOne"
- split: 0.2
+ split: 0.9
nb_folds: 2
nb_class: 2
classes:
- type: ["monoview", "multiview"]
- algos_monoview: ["decision_tree", "adaboost", "svm_linear", "random_forest"]
- algos_multiview: ["weighted_linear_early_fusion", "difficulty_fusion", "double_fault_fusion"]
- stats_iter: 30
+ type: ["multiview", "monoview"]
+ algos_monoview: ["decision_tree", "adaboost", "random_forest" ]
+ algos_multiview: ["weighted_linear_early_fusion",]
+ stats_iter: 1
metrics: ["accuracy_score", "f1_score"]
- metric_princ: "accuracy_score"
+ metric_princ: "f1_score"
hps_type: "randomized_search-equiv"
hps_iter: 5
@@ -65,7 +65,7 @@ adaboost_graalpy:
n_stumps: [1]
decision_tree:
- max_depth: [10]
+ max_depth: [2]
criterion: ["gini"]
splitter: ["best"]
diff --git a/multiview_platform/mono_multi_view_classifiers/multiview_classifiers/weighted_linear_early_fusion.py b/multiview_platform/mono_multi_view_classifiers/multiview_classifiers/weighted_linear_early_fusion.py
index 484b6657..170c1010 100644
--- a/multiview_platform/mono_multi_view_classifiers/multiview_classifiers/weighted_linear_early_fusion.py
+++ b/multiview_platform/mono_multi_view_classifiers/multiview_classifiers/weighted_linear_early_fusion.py
@@ -34,7 +34,7 @@ class WeightedLinearEarlyFusion(BaseMultiviewClassifier, BaseFusionClassifier):
super(WeightedLinearEarlyFusion, self).__init__(random_state=random_state)
self.view_weights = view_weights
self.monoview_classifier_name = monoview_classifier_name
- self.short_name = "early fusion " + monoview_classifier_name
+ self.short_name = "early fusion " + self.monoview_classifier_name
if monoview_classifier_name in monoview_classifier_config:
self.monoview_classifier_config = monoview_classifier_config[monoview_classifier_name]
self.monoview_classifier_config = monoview_classifier_config
@@ -59,6 +59,7 @@ class WeightedLinearEarlyFusion(BaseMultiviewClassifier, BaseFusionClassifier):
self.monoview_classifier = monoview_classifier_class()
self.init_monoview_estimator(monoview_classifier_name,
monoview_classifier_config)
+ self.short_name = "early fusion " + self.monoview_classifier_name
return self
def get_params(self, deep=True):
diff --git a/multiview_platform/mono_multi_view_classifiers/result_analysis.py b/multiview_platform/mono_multi_view_classifiers/result_analysis.py
index f529f1af..fdf0928d 100644
--- a/multiview_platform/mono_multi_view_classifiers/result_analysis.py
+++ b/multiview_platform/mono_multi_view_classifiers/result_analysis.py
@@ -4,6 +4,7 @@ import logging
import os
import time
import yaml
+import traceback
import matplotlib as mpl
from matplotlib.patches import Patch
@@ -162,6 +163,8 @@ def plot_metric_scores(train_scores, test_scores, names, nb_results, metric_name
))
fig.update_layout(title=metric_name + "\n" + tag + " scores for each classifier")
+ fig.update_layout(paper_bgcolor = 'rgba(0,0,0,0)',
+ plot_bgcolor = 'rgba(0,0,0,0)')
plotly.offline.plot(fig, filename=file_name + ".html", auto_open=False)
del fig
@@ -232,7 +235,8 @@ def plot_2d(data, classifiers_names, nbClassifiers, nbExamples,
reversescale=True), row=row_index+1, col=1)
fig.update_yaxes(title_text="Label "+str(row_index), showticklabels=False, ticks='', row=row_index+1, col=1)
fig.update_xaxes(showticklabels=False, row=row_index+1, col=1)
-
+ fig.update_layout(paper_bgcolor = 'rgba(0,0,0,0)',
+ plot_bgcolor = 'rgba(0,0,0,0)')
fig.update_xaxes(showticklabels=True, row=len(label_index_list), col=1)
plotly.offline.plot(fig, filename=file_name + "error_analysis_2D.html", auto_open=False)
del fig
@@ -629,6 +633,8 @@ def publish_feature_importances(feature_importances, directory, database_name, l
fig.update_layout(
xaxis={"showgrid": False, "showticklabels": False, "ticks": ''},
yaxis={"showgrid": False, "showticklabels": False, "ticks": ''})
+ fig.update_layout(paper_bgcolor = 'rgba(0,0,0,0)',
+ plot_bgcolor = 'rgba(0,0,0,0)')
plotly.offline.plot(fig, filename=file_name + ".html", auto_open=False)
del fig
@@ -724,7 +730,7 @@ def analyze_biclass(results, benchmark_argument_dictionaries, stats_iter, metric
logging.debug("Srart:\t Analzing all biclass resuls")
biclass_results = {}
flagged_tracebacks_list = []
-
+ fig_errors = []
for flag, result, tracebacks in results:
iteridex, [classifierPositive, classifierNegative] = flag
@@ -739,14 +745,13 @@ def analyze_biclass(results, benchmark_argument_dictionaries, stats_iter, metric
labels_names = [arguments["labels_dictionary"][0],
arguments["labels_dictionary"][1]]
+ flagged_tracebacks_list += publish_tracebacks(directory, database_name, labels_names, tracebacks, flag)
results = publishMetricsGraphs(metrics_scores, directory, database_name,
- labels_names)
+ labels_names)
publishExampleErrors(example_errors, directory, database_name,
labels_names, example_ids, arguments["labels"])
publish_feature_importances(feature_importances, directory, database_name, labels_names)
- flagged_tracebacks_list += publish_tracebacks(directory, database_name, labels_names, tracebacks, flag)
-
if not str(classifierPositive) + str(classifierNegative) in biclass_results:
biclass_results[str(classifierPositive) + str(classifierNegative)] = {}
diff --git a/multiview_platform/mono_multi_view_classifiers/utils/hyper_parameter_search.py b/multiview_platform/mono_multi_view_classifiers/utils/hyper_parameter_search.py
index bc27b3d3..46617c6e 100644
--- a/multiview_platform/mono_multi_view_classifiers/utils/hyper_parameter_search.py
+++ b/multiview_platform/mono_multi_view_classifiers/utils/hyper_parameter_search.py
@@ -180,12 +180,12 @@ class MultiviewCompatibleRandomizedSearchCV(RandomizedSearchCV):
def fit_multiview(self, X, y=None, groups=None, **fit_params):
n_splits = self.cv.get_n_splits(self.available_indices, y[self.available_indices])
- folds = self.cv.split(self.available_indices, y[self.available_indices])
+ folds = list(self.cv.split(self.available_indices, y[self.available_indices]))
+ if self.equivalent_draws:
+ self.n_iter = self.n_iter*X.nb_view
candidate_params = list(self._get_param_iterator())
base_estimator = clone(self.estimator)
results = {}
- if self.equivalent_draws:
- self.n_iter = self.n_iter*X.nb_view
self.cv_results_ = dict(("param_"+param_name, []) for param_name in candidate_params[0].keys())
self.cv_results_["mean_test_score"] = []
for candidate_param_idx, candidate_param in enumerate(candidate_params):
--
GitLab