Parcourir la source

use of openML problems for feature selection

Jérôme BUISINE il y a 3 ans
Parent
commit
846fea6b62

Fichier diff supprimé car celui-ci est trop grand
+ 3752 - 0
OpenML_datasets/bio_response.csv


Fichier diff supprimé car celui-ci est trop grand
+ 3470 - 0
OpenML_datasets/gina_agnostic.csv


Fichier diff supprimé car celui-ci est trop grand
+ 4231 - 0
OpenML_datasets/hiva_agnostic.csv


Fichier diff supprimé car celui-ci est trop grand
+ 2601 - 0
OpenML_datasets/madelon.csv


+ 244 - 0
find_best_attributes_surrogate_openML.py

@@ -0,0 +1,244 @@
+# main imports
+import os
+import sys
+import argparse
+import pandas as pd
+import numpy as np
+import logging
+import datetime
+import random
+
+# model imports
+from sklearn.model_selection import train_test_split
+from sklearn.model_selection import GridSearchCV
+from sklearn.linear_model import LogisticRegression
+from sklearn.ensemble import RandomForestClassifier, VotingClassifier
+
+from keras.layers import Dense, Dropout, LSTM, Embedding, GRU, BatchNormalization
+from keras.preprocessing.sequence import pad_sequences
+from keras.models import Sequential
+
+import joblib
+import sklearn
+import sklearn.svm as svm
+from sklearn.utils import shuffle
+from sklearn.metrics import roc_auc_score
+from sklearn.model_selection import cross_val_score
+from sklearn.preprocessing import MinMaxScaler
+
+# modules and config imports
+sys.path.insert(0, '') # trick to enable import of main folder module
+
+import custom_config as cfg
+import models as mdl
+
+from optimization.ILSSurrogate import ILSSurrogate
+from macop.solutions.BinarySolution import BinarySolution
+
+from macop.operators.mutators.SimpleMutation import SimpleMutation
+from macop.operators.mutators.SimpleBinaryMutation import SimpleBinaryMutation
+from macop.operators.crossovers.SimpleCrossover import SimpleCrossover
+from macop.operators.crossovers.RandomSplitCrossover import RandomSplitCrossover
+
+from macop.operators.policies.UCBPolicy import UCBPolicy
+
+from macop.callbacks.BasicCheckpoint import BasicCheckpoint
+from macop.callbacks.UCBCheckpoint import UCBCheckpoint
+
+from sklearn.ensemble import RandomForestClassifier
+
+
+# default validator
+def validator(solution):
+
+    # at least 5 attributes
+    if list(solution.data).count(1) < 5:
+        return False
+
+    return True
+
+def train_model(X_train, y_train):
+
+    print ('Creating model...')
+    # here use of SVM with grid search CV
+    Cs = [0.001, 0.01, 0.1, 1, 10, 100, 1000]
+    gammas = [0.001, 0.01, 0.1, 5, 10, 100]
+    param_grid = {'kernel':['rbf'], 'C': Cs, 'gamma' : gammas}
+
+    svc = svm.SVC(probability=True, class_weight='balanced')
+    #clf = GridSearchCV(svc, param_grid, cv=5, verbose=1, scoring=my_accuracy_scorer, n_jobs=-1)
+    clf = GridSearchCV(svc, param_grid, cv=5, verbose=1, n_jobs=-1)
+
+    clf.fit(X_train, y_train)
+
+    model = clf.best_estimator_
+
+    return model
+
+def loadDataset(filename):
+
+    # TODO : load data using DL RNN 
+
+    ########################
+    # 1. Get and prepare data
+    ########################
+    dataset = pd.read_csv(filename, sep=',')
+
+    # change label as common
+    min_label_value = min(dataset.iloc[:, -1])
+    max_label_value = max(dataset.iloc[:, -1])
+
+    dataset.iloc[:, -1] = dataset.iloc[:, -1].replace(min_label_value, 0)
+    dataset.iloc[:, -1] = dataset.iloc[:, -1].replace(max_label_value, 1)
+
+    X_dataset = dataset.iloc[:, :-1]
+    y_dataset = dataset.iloc[:, -1]
+
+    problem_size = len(X_dataset.columns)
+
+    # min/max normalisation over feature
+    # create a scaler object
+    scaler = MinMaxScaler()
+    # fit and transform the data
+    X_dataset = np.array(pd.DataFrame(scaler.fit_transform(X_dataset), columns=X_dataset.columns))
+
+    # prepare train, validation and test datasets
+    X_train, X_test, y_train, y_test = train_test_split(X_dataset, y_dataset, test_size=0.3, shuffle=True)
+
+    return X_train, y_train, X_test, y_test, problem_size
+
+
+def main():
+
+    parser = argparse.ArgumentParser(description="Train and find best filters to use for model")
+
+    parser.add_argument('--data', type=str, help='open ml dataset filename prefix', required=True)
+    #parser.add_argument('--start_surrogate', type=int, help='number of evalution before starting surrogare model', default=100)
+    parser.add_argument('--ils', type=int, help='number of total iteration for ils algorithm', required=True)
+    parser.add_argument('--ls', type=int, help='number of iteration for Local Search algorithm', required=True)
+    parser.add_argument('--output', type=str, help='output surrogate model name')
+
+    args = parser.parse_args()
+
+    p_data_file = args.data
+    #p_start     = args.start_surrogate
+    p_ils_iteration = args.ils
+    p_ls_iteration  = args.ls
+    p_output = args.output
+
+    # load data from file and get problem size
+    X_train, y_train, X_test, y_test, problem_size = loadDataset(p_data_file)
+
+    # create `logs` folder if necessary
+    if not os.path.exists(cfg.output_logs_folder):
+        os.makedirs(cfg.output_logs_folder)
+
+    logging.basicConfig(format='%(asctime)s %(message)s', filename='data/logs/{0}.log'.format(p_output), level=logging.DEBUG)
+
+    # init solution (`n` attributes)
+    def init():
+        return BinarySolution([], problem_size).random(validator)
+
+    # define evaluate function here (need of data information)
+    def evaluate(solution):
+
+        start = datetime.datetime.now()
+
+        # get indices of filters data to use (filters selection from solution)
+        indices = []
+
+        for index, value in enumerate(solution.data): 
+            if value == 1: 
+                indices.append(index) 
+
+        # keep only selected filters from solution
+        x_train_filters = X_train[:, indices]
+        x_test_filters = X_test[ :, indices]
+        
+        # model = mdl.get_trained_model(p_choice, x_train_filters, y_train_filters)
+        model = train_model(x_train_filters, y_train)
+
+        y_test_model = model.predict(x_test_filters)
+        y_test_predict = [ 1 if x > 0.5 else 0 for x in y_test_model ]
+        test_roc_auc = roc_auc_score(y_test, y_test_predict)
+
+        end = datetime.datetime.now()
+
+        diff = end - start
+
+        print("Real evaluation took: {}, score found: {}".format(divmod(diff.days * 86400 + diff.seconds, 60), test_roc_auc))
+
+        return test_roc_auc
+
+
+    # build all output folder and files based on `output` name
+    backup_model_folder = os.path.join(cfg.output_backup_folder, p_output)
+    surrogate_output_model = os.path.join(cfg.output_surrogates_model_folder, p_output)
+    surrogate_output_data = os.path.join(cfg.output_surrogates_data_folder, p_output)
+
+    if not os.path.exists(backup_model_folder):
+        os.makedirs(backup_model_folder)
+
+    if not os.path.exists(cfg.output_surrogates_model_folder):
+        os.makedirs(cfg.output_surrogates_model_folder)
+
+    if not os.path.exists(cfg.output_surrogates_data_folder):
+        os.makedirs(cfg.output_surrogates_data_folder)
+
+    backup_file_path = os.path.join(backup_model_folder, p_output + '.csv')
+    ucb_backup_file_path = os.path.join(backup_model_folder, p_output + '_ucbPolicy.csv')
+
+    # prepare optimization algorithm (only use of mutation as only ILS are used here, and local search need only local permutation)
+    operators = [SimpleBinaryMutation(), SimpleMutation()]
+    policy = UCBPolicy(operators)
+
+    # define first line if necessary
+    if not os.path.exists(surrogate_output_data):
+        folder, _ = os.path.split(surrogate_output_data)
+
+        if not os.path.exists(folder):
+            os.makedirs(folder)
+
+        with open(surrogate_output_data, 'w') as f:
+            f.write('x;y\n')
+
+
+    # custom start surrogate variable based on problem size
+    p_start = int(problem_size)
+    print(f'Starting using surrogate after {p_start} reals training')
+
+    # custom ILS for surrogate use
+    algo = ILSSurrogate(_initalizer=init, 
+                        _evaluator=evaluate, # same evaluator by defadefaultult, as we will use the surrogate function
+                        _operators=operators, 
+                        _policy=policy, 
+                        _validator=validator,
+                        _surrogate_file_path=surrogate_output_model,
+                        _start_train_surrogate=p_start, # start learning and using surrogate after 1000 real evaluation
+                        _solutions_file=surrogate_output_data,
+                        _ls_train_surrogate=1,
+                        _maximise=True)
+    
+    algo.addCallback(BasicCheckpoint(_every=1, _filepath=backup_file_path))
+    algo.addCallback(UCBCheckpoint(_every=1, _filepath=ucb_backup_file_path))
+
+    bestSol = algo.run(p_ils_iteration, p_ls_iteration)
+
+    # print best solution found
+    print("Found ", bestSol)
+
+    # save model information into .csv file
+    if not os.path.exists(cfg.results_information_folder):
+        os.makedirs(cfg.results_information_folder)
+
+    filename_path = os.path.join(cfg.results_information_folder, cfg.optimization_attributes_result_filename)
+
+    line_info = p_data_file + ';' + str(p_ils_iteration) + ';' + str(p_ls_iteration) + ';' + str(bestSol.data) + ';' + str(list(bestSol.data).count(1)) + ';' + str(bestSol.fitness())
+    with open(filename_path, 'a') as f:
+        f.write(line_info + '\n')
+    
+    print('Result saved into %s' % filename_path)
+
+
+if __name__ == "__main__":
+    main()

+ 10 - 10
optimization/LSSurrogate.py

@@ -22,19 +22,19 @@ class LocalSearchSurrogate(Algorithm):
         bestSolution: {Solution} -- best solution found so far during running algorithm
         callbacks: {[Callback]} -- list of Callback class implementation to do some instructions every number of evaluations and `load` when initializing algorithm
     """
-    def run(self, _evaluations):
+    def run(self, evaluations):
         """
         Run the local search algorithm
 
         Args:
-            _evaluations: {int} -- number of Local search evaluations
+            evaluations: {int} -- number of Local search evaluations
             
         Returns:
             {Solution} -- best solution found
         """
 
         # by default use of mother method to initialize variables
-        super().run(_evaluations)
+        super().run(evaluations)
 
         # do not use here the best solution known (default use of initRun and current solution)
         # if self.parent:
@@ -66,8 +66,8 @@ class LocalSearchSurrogate(Algorithm):
                              (newSolution, newSolution.fitness()))
 
                 # add to surrogate pool file if necessary (using ILS parent reference)
-                if self.parent.start_train_surrogate >= self.getGlobalEvaluation():
-                    self.parent.add_to_surrogate(newSolution)
+                # if self.parent.start_train_surrogate >= self.getGlobalEvaluation():
+                #     self.parent.add_to_surrogate(newSolution)
 
                 # stop algorithm if necessary
                 if self.stop():
@@ -81,17 +81,17 @@ class LocalSearchSurrogate(Algorithm):
 
         return self.bestSolution
 
-    def addCallback(self, _callback):
+    def addCallback(self, callback):
         """Add new callback to algorithm specifying usefull parameters
 
         Args:
-            _callback: {Callback} -- specific Callback instance
+            callback: {Callback} -- specific Callback instance
         """
         # specify current main algorithm reference
         if self.parent is not None:
-            _callback.setAlgo(self.parent)
+            callback.setAlgo(self.parent)
         else:
-            _callback.setAlgo(self)
+            callback.setAlgo(self)
 
         # set as new
-        self.callbacks.append(_callback)
+        self.callbacks.append(callback)

+ 35 - 0
run_openML_surrogate.py

@@ -0,0 +1,35 @@
+import os, argparse
+
+open_ml_problems_folder = 'OpenML_datasets'
+
+def main():
+
+    parser = argparse.ArgumentParser(description="Find best features for each OpenML problems")
+
+    parser.add_argument('--ils', type=int, help='number of total iteration for ils algorithm', required=True)
+    parser.add_argument('--ls', type=int, help='number of iteration for Local Search algorithm', required=True)
+
+    args = parser.parse_args()
+
+    p_ils = args.ils
+    p_ls  = args.ls
+
+    open_ml_problems = os.listdir(open_ml_problems_folder)
+
+    for ml_problem in open_ml_problems:
+
+        ml_problem_name = ml_problem.replace('.csv', '')
+        ml_problem_path = os.path.join(open_ml_problems_folder, ml_problem)
+
+        ml_surrogate_command = f"python find_best_attributes_surrogate_openML.py " \
+                               f"--data {ml_problem_path} " \
+                               f"--ils {p_ils} " \
+                               f"--ls {p_ls} " \
+                               f"--output {ml_problem_name}"
+        print(f'Run surrogate features selection for {ml_problem_name} with [ils: {p_ils}, ls: {p_ls}]')
+        print(ml_surrogate_command)
+        os.system(ml_surrogate_command)
+    
+
+if __name__ == "__main__":
+    main()