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
|
import pandas as pd
from sklearn.model_selection import StratifiedKFold
from pipeline.ngrams_classif import NgramsExtractor
from sklearn.pipeline import FeatureUnion, Pipeline
from sklearn.ensemble import RandomForestClassifier
from sklearn.metrics import accuracy_score, recall_score, f1_score, precision_score
import numpy as np
import warnings
warnings.filterwarnings('ignore')
def average(arr: list):
return sum(arr) / len(arr)
def classify_ndss(train, test):
# print(test)
combinedFeatures = FeatureUnion([
# ('tsfresh', TSFreshBasicExtractor()),
('ngrams', NgramsExtractor(max_ngram_len=1)),
])
pipeline = Pipeline([
('features', combinedFeatures),
('clf', RandomForestClassifier(n_estimators=100, max_depth=30, min_samples_leaf=5)),
# ('clf', RandomForestClassifier(n_estimators=100)),
])
pipeline.fit(train, list(train.class_label))
# Prediction
pred_ret = pipeline.predict(test)
return pred_ret
def classify_ndss_key(train, test):
# print(test)
combinedFeatures = FeatureUnion([
# ('tsfresh', TSFreshBasicExtractor()),
('ngrams', NgramsExtractor(max_ngram_len=1)),
])
pipeline = Pipeline([
('features', combinedFeatures),
('clf', RandomForestClassifier(n_estimators=100, max_depth=30, min_samples_leaf=5)),
# ('clf', RandomForestClassifier(n_estimators=100)),
])
pipeline.fit(train, list(train.class_label))
# Prediction
pred_ret = pipeline.predict(test)
return pred_ret
def classify_rf(train, test):
rf = RandomForestClassifier()
X = []
Y = []
for index, row in train.iterrows():
lengths = list(row['lengths'])
if len(lengths) > 100:
lengths = lengths[:100]
elif len(lengths) < 100:
while len(lengths) < 100:
lengths.append(0)
X.append(lengths)
label = int(row['class_label'])
Y.append(label)
rf.fit(X, Y)
# return
test_X = []
for index, row in test.iterrows():
lengths = list(row['lengths'])
if len(lengths) > 100:
lengths = lengths[:100]
elif len(lengths) < 100:
while len(lengths) < 100:
lengths.append(0)
test_X.append(lengths)
pred_ret = rf.predict(test_X)
return pred_ret
def trans_csv_to_df(csv_filename):
src_df = pd.read_csv(csv_filename)
dst_df = pd.DataFrame()
for i in range(len(src_df)):
features = np.array(eval(src_df.loc[i, 'lengths']))
label = src_df.loc[i, 'class_label']
# print(label,type(label))
if label >= 100:
continue
# dst_df.
dst_df = dst_df.append({
"lengths": features,
"class_label": label
}, ignore_index=True)
return dst_df
def exp(classifier, feature_model, data_model, ops_mode="win10"):
if classifier == "rf":
classify = classify_rf
elif classifier == "ndss":
if feature_model == "norm":
classify = classify_ndss
elif feature_model == "key":
classify = classify_ndss_key
else:
print("feature_model", feature_model)
return
else:
print("未知classifier", classifier)
return
if feature_model in ["key", "norm"]:
df = trans_csv_to_df(f"./data/{feature_model}_feature_{data_model}_{ops_mode}.csv")
else:
print("未知特征类别！")
return
kf = StratifiedKFold(n_splits=10, shuffle=True)
precisions = []
recalls = []
f1s = []
accs = []
for k, (train, test) in enumerate(kf.split(df, list(df.class_label))):
if classifier == "ndss":
predict_results = classify(df.iloc[train], df.iloc[test])
elif classifier == "rf":
predict_results = classify(df.iloc[train], df.iloc[test])
else:
print("未知分类方法")
return
gt_Y = df.iloc[test].class_label
precision = precision_score(gt_Y, predict_results, average='weighted')
recall = recall_score(gt_Y, predict_results, average='weighted')
f1 = f1_score(gt_Y, predict_results, average='weighted')
acc = accuracy_score(gt_Y, predict_results)
precisions.append(precision)
recalls.append(recall)
f1s.append(f1)
accs.append(acc)
break
print("平均准确率: ", average(precisions), end="\t")
print("平均召回率: ", average(recalls), end="\t")
print("平均f1值: ", average(f1s), end="\t")
print("平均acc: ", average(accs))
def cross_validation(classifier, feature_model):
if classifier == "rf":
classify = classify_rf
elif classifier == "ndss":
if feature_model == "norm":
classify = classify_ndss
elif feature_model == "key":
classify = classify_ndss_key
else:
print("feature_model", feature_model)
return
else:
print("未知classifier", classifier)
return
# train = "firefox"
# test = "chrome"
print("classifier:", classifier)
print("feature_model", feature_model)
# for train, test in [("chrome", "edge"), ("chrome", "firefox"), ("firefox", "chrome"), ("firefox", "edge"),
# ("edge", "chrome"), ("edge", "firefox")]:
# print("train:", train, "test:", test)
# df_train = trans_csv_to_df(f"./data/{feature_model}_feature_{train}_win10.csv")
# df_test = trans_csv_to_df(f"./data/{feature_model}_feature_{test}_win10.csv")
# predict_results = classify(df_train, df_test)
# gt_Y = df_test.class_label
# precision = precision_score(gt_Y, predict_results, average='weighted')
# recall = recall_score(gt_Y, predict_results, average='weighted')
# f1 = f1_score(gt_Y, predict_results, average='weighted')
# acc = accuracy_score(gt_Y, predict_results)
# print("准确率: ", precision, end="\t")
# print("召回率: ", recall, end="\t")
# print("f1值: ", f1, end="\t")
# print("acc: ", acc)
for train, test in [("win10", "ubuntu"), ("ubuntu", "win10")]:
print("train:", train, "test:", test)
df_train = trans_csv_to_df(f"./data/{feature_model}_feature_chrome_{train}.csv")
df_test = trans_csv_to_df(f"./data/{feature_model}_feature_chrome_{test}.csv")
predict_results = classify(df_train, df_test)
gt_Y = df_test.class_label
precision = precision_score(gt_Y, predict_results, average='weighted')
recall = recall_score(gt_Y, predict_results, average='weighted')
f1 = f1_score(gt_Y, predict_results, average='weighted')
acc = accuracy_score(gt_Y, predict_results)
print("准确率: ", precision, end="\t")
print("召回率: ", recall, end="\t")
print("f1值: ", f1, end="\t")
print("acc: ", acc)
if __name__ == '__main__':
exp("ndss", "norm", "firefox")
exp("ndss", "key", "firefox")
#
exp("ndss", "norm", "chrome")
exp("ndss", "key", "chrome")
exp("ndss", "norm", "chrome", "ubuntu")
exp("ndss", "key", "chrome", "ubuntu")
cross_validation("ndss", "norm")
cross_validation("ndss", "key")
|
