-
Notifications
You must be signed in to change notification settings - Fork 0
/
Copy pathtest_mce.py
155 lines (107 loc) · 4.8 KB
/
test_mce.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
from sklearn.ensemble import RandomForestClassifier
from sklearn.neural_network import MLPClassifier
from sklearn.model_selection import train_test_split
from sklearn.metrics import accuracy_score
import joblib
from pathlib import Path
from explainer import DefaultExplainer
from visualizer import ExplanationVisualizer
from data_loader import load_data_txt, load_heloc, load_kdd_csv, load_ids_csv
import utils
import numpy as np
def test():
chosen_attributes = [0, 5]
clf = RandomForestClassifier(n_jobs=100, n_estimators=100, random_state=5000)
X, Y, names = load_data_txt(normalize=True)
X_train, X_test, Y_train, Y_test = train_test_split(X, Y, test_size=0.2, random_state=1001)
clf.fit(X_train, Y_train)
explainer = DefaultExplainer(clf, X, [0, 5])
for instance in X_test:
if clf.predict(instance.reshape(1, -1)) < 0.5:
explainer.explain_instance(instance)
break
y_clf = clf.predict(X_test)
y_exp = explainer.sg.surrogate.predict(X_test)
print('comparison score on test dataset: ', accuracy_score(y_clf, y_exp))
viz = ExplanationVisualizer(explainer, chosen_attributes, feature_names=names)
viz.present_explanation('relative')
def test_heloc():
clf = RandomForestClassifier(n_jobs=100, n_estimators=100, random_state=1000)
X, Y, names = load_heloc(normalize=False)
X_train, X_test, Y_train, Y_test = train_test_split(X, Y, test_size=0.2, random_state=1200)
clf.fit(X_train, Y_train)
print('accuracy Rf: ', accuracy_score(clf.predict(X_test), Y_test))
explainer = DefaultExplainer(clf, X)
for instance in X_test:
if clf.predict_proba(instance.reshape(1, -1))[0, 1] < 0.5:
explainer.explain_instance(instance)
break
viz = ExplanationVisualizer(explainer, feature_names=names)
viz.present_explanation(method='relative')
def runon_kdd():
mlp_dump_file = "exports/mlp.joblib"
p = Path(mlp_dump_file)
X, Y, names= load_kdd_csv(normalize=True, train=True)
if p.is_file():
clf = joblib.load(mlp_dump_file)
else:
clf = MLPClassifier(solver='adam', alpha=1e-2, hidden_layer_sizes = (20, 5), random_state = 1)
clf.fit(X, Y)
joblib.dump(clf, mlp_dump_file)
Xtest, Ytest, names = load_kdd_csv(normalize=True, train=False)
print('accuracy MLP: ', accuracy_score(clf.predict(Xtest), Ytest))
explainer = DefaultExplainer(clf, X, None)
for instance in Xtest:
if clf.predict(instance.reshape(1, -1)) < 0.5:
explainer.explain_instance(instance)
break
print('counterfact: ', clf.predict(explainer.counterfactual.reshape(1, -1)))
viz = ExplanationVisualizer(explainer, None, feature_names=names)
viz.present_explanation(method='relative')
def runon_ids():
mlp_dump_file = "exports/mlp_ids.joblib"
p = Path(mlp_dump_file)
X, Y, names = load_ids_csv(normalize=True)
choose = np.random.randint(X.shape[0], size=50)
X_train, X_test, Y_train, Y_test = train_test_split(X, Y, test_size=0.2, random_state=1203)
if p.is_file():
clf = joblib.load(mlp_dump_file)
else:
clf = MLPClassifier(solver='adam', alpha=1e-2, hidden_layer_sizes = (20, 5), random_state = 1)
clf.fit(X_train, Y_train)
joblib.dump(clf, mlp_dump_file)
pred = clf.predict(X_test)
false = X_test[pred != Y_test]
explainer = DefaultExplainer(clf, X, features_names=names)
global_instance = 0
for instance in X_test:
if clf.predict(instance.reshape(1, -1)) < 0.5: # explain attacks (0)
explainer.explain_instance(instance)
global_instance = instance
break
print('accuracy MLP: ', accuracy_score(clf.predict(X_test), Y_test))
viz = ExplanationVisualizer(explainer, None, feature_names=names)
viz.present_explanation(method='relative')
# viz.present_explanation(method='visual')
def tsne_on_ids():
mlp_dump_file = "exports/mlp_ids.joblib"
p = Path(mlp_dump_file)
X, Y, names = load_ids_csv(normalize=True)
X_train, X_test, Y_train, Y_test = train_test_split(X, Y, test_size=0.2, random_state=1201)
if p.is_file():
clf = joblib.load(mlp_dump_file)
else:
clf = MLPClassifier(solver='adam', alpha=1e-2, hidden_layer_sizes = (20, 5), random_state = 1)
clf.fit(X_train, Y_train)
joblib.dump(clf, mlp_dump_file)
choose = np.random.randint(X.shape[0], size=50)
data_subset = X[choose, :]
label_subset = Y[choose]
explainer = DefaultExplainer(clf, X, None)
counterfacts = explainer.get_counterfactuals(data_subset[label_subset == 0])
distances = abs(counterfacts - data_subset[label_subset == 0])
print(np.average(distances, axis=0))
utils.plot_tsne(data_subset, label_subset, counterfacts)
return
if __name__ == '__main__':
runon_ids()