-
Notifications
You must be signed in to change notification settings - Fork 0
/
hyperOp_bln.py
217 lines (205 loc) · 10.3 KB
/
hyperOp_bln.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
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
from sklearn import metrics
import warnings
warnings.filterwarnings('ignore')
from sklearn.exceptions import FitFailedWarning
from sklearn.model_selection import GridSearchCV
from sklearn.model_selection import train_test_split
from sklearn.metrics import classification_report
import numpy as np
from sklearn.model_selection import RepeatedKFold, train_test_split
import pickle
import os
import pandas as pd
from sklearn import metrics
import time
from RandomForests.RF import doRF
from sklearn.ensemble import RandomForestClassifier
from NaiveBayes.NB import doNB
from sklearn.naive_bayes import GaussianNB
from CART.docart import doCART
from sklearn.tree import DecisionTreeClassifier
from LR.dolr import doLR
from sklearn.linear_model import LogisticRegression
from KNN.simple_doknn import doKNN
from sklearn.neighbors import KNeighborsClassifier
# from SVM.SVM import doSVM
# from sklearn import svm
# from extensions.ANN.doann import doANN
# from sklearn.neural_network import MLPClassifier
# from Kmeans.Kmeans import doKmeans
from PCA.doPCA import applyPCA, getData, CSV_TARGET_COLUMNS
from PCA.doKPCA import applyKPCA
import pickle
import os
PRINCIPLE_COMPONENT_FINDER = applyPCA
PROJECTS = [ 'Mirantis', 'Wikimedia','Mozilla','Openstack']
score = 'balanced_accuracy'
# score = 'balanced_accuracy'
ALGORITHMS_NAME=[doLR,doCART,doKNN,doNB,doRF]
ALGORITHMS = {
doCART.__name__: DecisionTreeClassifier,
doNB.__name__: GaussianNB,
doKNN.__name__: KNeighborsClassifier,
doLR.__name__: LogisticRegression,
doRF.__name__: RandomForestClassifier,
}
appDirectory='.'
step=20
for algo in ALGORITHMS_NAME:
for project in PROJECTS:
precision_score = []
scores_list = []
auc_score = []
recall_score = []
f1_score = []
precision_score_no_optim = []
scores_list_no_optim = []
auc_score_no_optim = []
recall_score_no_optim = []
f1_score_no_optim = []
best_params=[];
for boostrap_index in range(0,100):
# bootstraping
startTime=time.time()
print('bootstrap round:',boostrap_index)
raw_data = getData(project,appDirectory);
# print(type(raw_data))
train_data_indexes = np.random.choice(raw_data.shape[0],raw_data.shape[0])
# print(train_data_indexes)
train_data = raw_data.loc[train_data_indexes]
train_data.reset_index(inplace=True)
test_data = raw_data.drop(np.unique(train_data_indexes))
test_data.reset_index(inplace=True)
if test_data.shape[0]==0:
continue
#finish bootstraping
print("# Tuning hyper-parameters->goal {} for {} and {} project".format(score,algo.__name__,project))
print()
algorithm = ALGORITHMS[algo.__name__]
dataSet,DimensionReductionModel = PRINCIPLE_COMPONENT_FINDER(rawData=train_data,appDirectory= appDirectory)
test_data_pc = DimensionReductionModel.transform(test_data.iloc[:,:12])
test_data_target = test_data[CSV_TARGET_COLUMNS].to_numpy()
print(np.sum(dataSet.target)/dataSet.target.shape[0])
# X_train=dataSet.components
# y_train= dataSet.target
max_iter = ['log2','sqrt',None]
GRID_VALUES={
doCART.__name__:{
'criterion':['gini','entropy'],
'splitter':['best', 'random'],
'min_samples_split':[2,5,8,10,100],
'min_samples_leaf':[1,3,6,10,100],
'min_weight_fraction_leaf':[0.0001, 0.001, 0.01, 0.1, 0.49],
'max_features': ['log2','sqrt',None],
# 'max_features': ['log2','sqrt',None,*list(range(1,dataSet.components.shape[1],2))],
},
doNB.__name__:{
},
doKNN.__name__:{
'n_neighbors':[*list(range(1,max(int(dataSet.components.shape[0]**(0.5)),17),4))], #
'weights':['uniform','distance'],
'algorithm':['auto', 'ball_tree', 'kd_tree', 'brute'],
'leaf_size':[*list(range(10,50,10))],
},
doLR.__name__:{
'tol':[1e-5,1e-4,1e-3,1e-2,1e-1], #
'C': [0.001, 0.1, 1, 10, 100], #
'fit_intercept':[True,False],
# 'solver':[ 'liblinear', 'saga'],
'penalty':[ 'l2' ],
'solver':[ 'lbfgs','sag','newton-cg'],
'max_iter':[50,100,150,200],
# 'l1_ratio':[1e-5,1e-3,1e-1,0.5],
# 'penalty':['elasticnet'],
# 'penalty':['l1', 'l2' ],
},
doRF.__name__:{
'n_estimators':[*list(range(10,50,10))], ##
'criterion':['gini','entropy'],
'min_samples_split':[2,5,8,10,100],
# 'min_samples_split':[2,4,6,8,10], #
'min_samples_leaf':[1,3,6,10,100],
# 'min_samples_leaf':[1,2,3,4,5,6], #
'min_weight_fraction_leaf':[0.0001, 0.001, 0.01, 0.1, 0.49],
'max_features': ['log2','sqrt',None],
}
}
# rkf = RepeatedKFold(n_splits=10, n_repeats=10, random_state=2652124)
# i=0
# X_train, X_test, y_train, y_test = train_test_split( dataSet.components, dataSet.target, test_size=0.2, random_state=1)
X_train = dataSet.components
y_train = dataSet.target
X_test = test_data_pc
y_test = test_data_target
clf = GridSearchCV(algorithm(), param_grid = GRID_VALUES[algo.__name__],scoring = score,n_jobs=-1,verbose=1,cv=10)
clf.fit(X_train, y_train)
print(clf.best_params_)
# for train_index, test_index in rkf.split(dataSet.components):
# i+=1
# print('round:',i)
# X_train, X_test, y_train, y_test = dataSet.components[train_index], dataSet.components[test_index], dataSet.target[train_index], dataSet.target[test_index]
# with warnings.catch_warnings():
# warnings.simplefilter("ignore",category=FitFailedWarning)
# try:
# clf.fit(X_train, y_train)
# except:
# continue
#testing with optimized params
y_pred = clf.predict(X_test)#, y_train, y_test,params = clf.best_params_)
try:
auc_score_op = metrics.roc_auc_score(y_test, y_pred)
except:
auc_score_op = 0
# print("optimized:",metrics.balanced_accuracy_score(y_test,y_pred))
print("optimized:",auc_score_op)
precision_score.append(metrics.precision_score(y_test, y_pred))
scores_list.append(metrics.accuracy_score(y_test, y_pred))
auc_score.append(auc_score_op)
recall_score.append(metrics.recall_score(y_test, y_pred))
f1_score.append(metrics.f1_score(y_test, y_pred))
# #testing with default params
result_no_optim = algo(X_train, X_test, y_train, y_test)
# print("not optimized:",result_no_optim.get('balanced_accuracy'))
print('not optimized:',result_no_optim.get('auc'))
precision_score_no_optim.append(result_no_optim.get('precision'))
scores_list_no_optim.append(result_no_optim.get('accuracy'))
auc_score_no_optim.append(result_no_optim.get('auc'))
recall_score_no_optim.append(result_no_optim.get('recall'))
f1_score_no_optim.append(result_no_optim.get('f1_measure'))
best_params.append({'params':clf.best_params_,
'elapsed_time': time.time() - startTime ,
'optim_res':{'precision':result_no_optim.get('precision'),'accuracy':result_no_optim.get('accuracy'),'auc':result_no_optim.get('auc'),'recall':result_no_optim.get('recall'),'f1_measure':result_no_optim.get('f1_measure')},
'defaul_res':{'precision':metrics.precision_score(y_test, y_pred),'accuracy':metrics.accuracy_score(y_test, y_pred),'auc':auc_score_op,'recall':metrics.recall_score(y_test, y_pred),'f1_measure':metrics.f1_score(y_test, y_pred)},
})
algorithm_final_result = {'precision:': np.median(precision_score),
'accuracy:': np.median(scores_list),
'auc:': np.median(auc_score),
'recall:': np.median(recall_score),
'f1_measure:': np.median(f1_score),
'auc_sd':np.std(auc_score),
'auc_avg':np.average(auc_score)
}
algorithm_final_result_no_optim = {'precision:': np.median(precision_score_no_optim),
'accuracy:': np.median(scores_list_no_optim),
'auc:': np.median(auc_score_no_optim),
'recall:': np.median(recall_score_no_optim),
'f1_measure:': np.median(f1_score_no_optim),
'auc_sd':np.std(auc_score_no_optim),
'auc_avg':np.average(auc_score_no_optim)
}
#saaving data
print("[{}] for {} is {}".format(algo.__name__, project, algorithm_final_result))
f = open('./Optimization_results/'+algo.__name__+'__'+project+'_params.pckl', 'wb')
pickle.dump(best_params, f)
f.close();
print(best_params)
f = open('./Optimization_results/'+algo.__name__+'__'+project+'.pckl', 'wb')
pickle.dump(algorithm_final_result, f)
f.close();
print("[{}] no-optim for {} is {}".format(algo.__name__, project, algorithm_final_result_no_optim))
f = open('./Optimization_results/'+algo.__name__+'_no_optim__'+project+'.pckl', 'wb')
pickle.dump(algorithm_final_result_no_optim, f)
f.close();
f = open('./Optimization_results/finished.pckl', 'wb')
pickle.dump("enjoy the results!", f)
f.close()