-
Notifications
You must be signed in to change notification settings - Fork 0
/
Copy path02.x.main_loop_C.py
191 lines (170 loc) · 8.45 KB
/
02.x.main_loop_C.py
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
import numpy
import time
from time import sleep
import scipy.io
from MachineSpecificSettings import Settings
from DataSetLoaderLib import DataSetLoader
from sklearn.neural_network import MLPClassifier
from sklearn.pipeline import make_pipeline
from sklearn.externals import joblib
from sklearn import preprocessing
from sklearn import svm
from sklearn import tree
#from sklearn.model_selection import cross_validation
from sklearn import metrics
from sklearn.ensemble import VotingClassifier
from sklearn.ensemble import AdaBoostClassifier
from sklearn.ensemble import ExtraTreesClassifier
from sklearn.ensemble import RandomForestClassifier
from sklearn.metrics import matthews_corrcoef,accuracy_score
from sklearn.model_selection import GridSearchCV
from sklearn.model_selection import train_test_split
#from sklearn.model_selection import LeaveOneOut
from sklearn.neural_network import MLPClassifier
#from sklearn.model_selection import StratifiedKFold
from sklearn.model_selection import cross_val_score
from sklearn.metrics import confusion_matrix
#import sklearn.cross_validation
'''
done - For each dataset, load the training and test dataset
done - Load the FSS as per method and size
done - for each of the classifers
done - split training set according to validation technique i.e. n_splits=10 for 10FoldCV and n for LOOCV
done - create a pipeline of classifier for both RAW and Normalized processing
done - fit & calculate the training accuracy on validation part of the dataset
done - test the 'fitted' classifier using the test dataset & record mcc value
done - record the complete details as a row in CSV
done - save the 'fitted' classifier for future reuse
'''
numpy.seterr(over='raise')
#Different feature selection methods
datasets=['C','B','A']
methods=['MRMR','JMI','JMIM']
sizes=['10','50','100','150','200','250']
classifiers = ["RandomForest","AdaBoost","DT","ExtraTree", "MLP","SVM"]
validationTechniques = ["10FoldCV"] #"LOOCV",
preps=["Standard","Robust","Quantile","Imputer"]
basePath='' #needed when we want to run it locally
#Iterating over each method
for dataset in datasets:
f=open('mcc/mccResults'+dataset+'.txt','a');
f.write('\n{date:%Y-%m-%d_%H:%M:%S}'.format( date=datetime.datetime.now() ))
#f.write("dataset, size, method, classifier, validationTechnique, mc, timeTaken, cv.max, cv.mean, cv.min, cv.std, preprocessing");
print "Dataset = ",dataset
#initiating datasetloader object
d = DataSetLoader();
#loading relevant Data and coresponding labels of dataset A
X_train_full = d.LoadDataSet(dataset+"_train");
y_train = d.LoadDataSetClasses(dataset+"_train");
X_validate_full = d.LoadDataSet(dataset+"_test");
y_validate = d.LoadDataSetClasses(dataset+"_test");
print ("Dimensions of training data and labels:",X_train_full.shape,y_train.shape)
print ("Dimensions of validation data and labels:",X_validate_full.shape,y_validate.shape)
#READY with Dataset, going to perform the main loop now
for method in methods:
#Iterating over each size
for size in sizes:
print ("Size and method:",size,method)
#first run indices
indices= joblib.load('datasetC_pickles/datasetC_train'+size+'-'+method+'.joblib.pkl')
X_train=X_train_full[:,indices]
X_validate=X_validate_full[:,indices]
for prepType in preps:
if prepType=="Standard":
preprocess = preprocessing.StandardScaler()
if prepType=="Robust":
preprocess = preprocessing.RobustScaler(with_centering=True, with_scaling=True, quantile_range=(25.0, 75.0), copy=True)
if prepType=="Quantile":
preprocess = preprocessing.QuantileTransformer(output_distribution='normal',n_quantiles=10, random_state=0)
if prepType=="Imputer":
preprocess = preprocessing.Imputer(missing_values='NaN', strategy='mean', axis=0)
for validation in validationTechniques:
if(validation=="LOOCV"):
validate = X_train_full.shape[0]-15
else:
validate = 10
for classifierName in classifiers:
start_time=time.time()
if(classifierName == "MLP"):
model = make_pipeline(
preprocess,
MLPClassifier(solver='adam',
alpha=0.0001,
activation='relu',
batch_size=150,
hidden_layer_sizes=(200, 100),
random_state=1))
classifier = make_pipeline(preprocess, MLPClassifier(activation='logistic',solver='sgd'))
# Construct the parameter grid
param_grid={
'mlpclassifier__learning_rate': ["constant", "invscaling", "adaptive"],
'mlpclassifier__alpha': [1, 0.1, 0.001, 0.0001],
'mlpclassifier__activation': ["logistic", "relu", "tanh"],
'mlpclassifier__hidden_layer_sizes': [(108,1), (108,2)],
'mlpclassifier__max_iter': [100, 500]
}
# Train the model
grid_clf = GridSearchCV(model,param_grid,cv=validate,iid=False)
classifier = grid_clf
if(classifierName == "SVM"): #kernel='linear'
model = make_pipeline(preprocess,svm.SVC( probability=True))
Cs = [0.001, 0.01, 0.1, 1, 10]
gammas = [0.001, 0.01, 0.1, 1]
kernels = ['rbf','linear']
param_grid = {'svc__C': Cs, 'svc__gamma' : gammas, 'svc__kernel':kernels}
grid_clf = GridSearchCV(model,param_grid,cv=validate,iid=False)
classifier = grid_clf
if(classifierName =="AdaBoost"):
classifier = make_pipeline(preprocess,AdaBoostClassifier())
if(classifierName =="DT"):
classifier = make_pipeline(preprocess,tree.DecisionTreeClassifier(max_depth=None, min_samples_split=2,random_state=0))
if(classifierName =="RandomForest"):
model = make_pipeline(preprocess,RandomForestClassifier())
param_grid = {
'randomforestclassifier__n_estimators': [100, 300, 500],
'randomforestclassifier__criterion': ['gini', 'entropy'],
'randomforestclassifier__bootstrap': [True, False]
}
grid_clf = GridSearchCV(model,param_grid,cv=validate,iid=False)
classifier = grid_clf
if(classifierName =="ExtraTree"):
classifier = make_pipeline(preprocess,ExtraTreesClassifier(n_estimators=10, max_depth=10000,min_samples_split=2, random_state=0))
classifier.fit(X_train,y_train)
extra_info = ""
for validationTechnique in validationTechniques:
if validationTechnique=="LOOCV":
#https://www.programcreek.com/python/example/91872/sklearn.cross_validation.LeaveOneOut
folds = cross_validation.LeaveOneOut(X_train.shape[0])
scores = cross_val_score(classifier, X_train, y_train, cv=folds, n_jobs=-1)
else:
folds = 10
scores = cross_val_score(classifier, X_train, y_train, cv=folds)
print scores.max(), scores.mean(), scores.min(), scores.std() * 2
if(classifierName == "MLP" or classifierName == "SVM" or classifierName=="RandomForest"):
y_pred = classifier.best_estimator_.predict(X_validate)
extra_info = classifier.best_params_
else:
y_pred = classifier.predict(X_validate);
print "------Validation Accuracy-------"
print y_pred.shape
print y_pred
print numpy.array(y_validate).shape
print y_validate
#transform data to be ready for mcc
y_pred=numpy.array(y_pred)
y_pred[y_pred == 0] = -1
y_validate=numpy.array(y_validate)
y_validate[y_validate == 0] = -1
mc = matthews_corrcoef(y_validate,y_pred);
cm = confusion_matrix (y_validate,y_pred);
"""
Add CM to the list of outputs
"""
end_time=time.time()-start_time
#now dump it in the file as CSV; dataset, size, method, classifier, validationTechnique, mc, timeTaken
f.write("\n "+dataset+", "+size+", "+method+", "+classifierName+", "+validationTechnique+", "+str(mc)+", "+str(end_time)+","+str(scores.max()) + "," + str(scores.mean()) + ", " + str(scores.min()) + ", " + str(scores.std() * 2)+", "+prepType);
print("matthew:",matthews_corrcoef(y_validate, y_pred))
print "Writing file: ", 'dataset'+dataset+'_trained_clfs/'+prepType+"-"+validationTechnique+"-"+classifierName+'-'+method+'-'+size+'.joblib.pkl'
joblib.dump(classifier,'dataset'+dataset+'_trained_clfs/'+prepType+"-"+validationTechnique+"-"+classifierName+'-'+method+'-'+size+'.joblib.pkl')
sleep(0.2);
f.close()