-
Notifications
You must be signed in to change notification settings - Fork 0
/
code.py
287 lines (287 loc) · 10.9 KB
/
code.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
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
import pandas as pd
import numpy as np
import sys
import sklearn
print(pd.__version__)
print(np.__version__)
print(sys.version)
print(sklearn.__version__)
from google.colab import files
uploaded = files.upload()
from google.colab import files
uploaded = files.upload()
col_names = ["duration","protocol_type","service","flag","src_bytes",
"dst_bytes","land","wrong_fragment","urgent","hot","num_failed_logins",
"logged_in","num_compromised","root_shell","su_attempted","num_root",
"num_file_creations","num_shells","num_access_files","num_outbound_cmds"
,
"is_host_login","is_guest_login","count","srv_count","serror_rate",
"srv_serror_rate","rerror_rate","srv_rerror_rate","same_srv_rate",
"diff_srv_rate","srv_diff_host_rate","dst_host_count","dst_host_srv_coun
t",
"dst_host_same_srv_rate","dst_host_diff_srv_rate","dst_host_same_src_por
t_rate",
"dst_host_srv_diff_host_rate","dst_host_serror_rate","dst_host_srv_serro
r_rate",
"dst_host_rerror_rate","dst_host_srv_rerror_rate","label"]
df = pd.read_csv("KDDTrain+_2.csv", header=None, names = col_names)
df_test = pd.read_csv("KDDTest+_2.csv", header=None, names = col_names)
print('Dimensions of the Training set:',df.shape)
print('Dimensions of the Test set:',df_test.shape)
df.head(5)
print('Label distribution Training set:')
print(df['label'].value_counts())
print()
print('Label distribution Test set:')
print(df_test['label'].value_counts())
service (column 3), flag (column 4).
print('Training set:')
for col_name in df.columns:
if df[col_name].dtypes == 'object' :
unique_cat = len(df[col_name].unique())
print("Feature '{col_name}' has {unique_cat} categories".format(col_
name=col_name, unique_cat=unique_cat))
erefore we need to make dummies for all.
print()
print('Distribution of categories in service:')
print(df['service'].value_counts().sort_values(ascending=False).head())
print('Test set:')
for col_name in df_test.columns:
if df_test[col_name].dtypes == 'object' :
unique_cat = len(df_test[col_name].unique())
print("Feature '{col_name}' has {unique_cat} categories".format(col_
name=col_name, unique_cat=unique_cat))
enc = OneHotEncoder()
df_categorical_values_encenc = enc.fit_transform(df_categorical_values_enc)
df_cat_data = pd.DataFrame(df_categorical_values_encenc.toarray(),columns=du
mcols)
testdf_categorical_values_encenc = enc.fit_transform(testdf_categorical_valu
es_enc)
testdf_cat_data = pd.DataFrame(testdf_categorical_values_encenc.toarray(),co
lumns=testdumcols)
df_cat_data.head()
trainservice=df['service'].tolist()
testservice= df_test['service'].tolist()
difference=list(set(trainservice) - set(testservice))
string = 'service_'
difference=[string + x for x in difference]
difference
for col in difference:
testdf_cat_data[col] = 0
testdf_cat_data.shape
newdf=df.join(df_cat_data)
newdf.drop('flag', axis=1, inplace=True)
newdf.drop('protocol_type', axis=1, inplace=True)
newdf.drop('service', axis=1, inplace=True)
newdf_test=df_test.join(testdf_cat_data)
newdf_test.drop('flag', axis=1, inplace=True)
newdf_test.drop('protocol_type', axis=1, inplace=True)
newdf_test.drop('service', axis=1, inplace=True)
print(newdf.shape)
print(newdf_test.shape)
labeldf=newdf['label']
labeldf_test=newdf_test['label']
newlabeldf=labeldf.replace({ 'normal' : 0, 'neptune' : 1 ,'back': 1, 'land':
1, 'pod': 1, 'smurf': 1, 'teardrop': 1,'mailbomb': 1, 'apache2': 1, 'proces
stable': 1, 'udpstorm': 1, 'worm': 1,
'ipsweep' : 2,'nmap' : 2,'portsweep' : 2,'satan'
: 2,'mscan' : 2,'saint' : 2
,'ftp_write': 3,'guess_passwd': 3,'imap': 3,'mult
ihop': 3,'phf': 3,'spy': 3,'warezclient': 3,'warezmaster': 3,'sendmail': 3,'
named': 3,'snmpgetattack': 3,'snmpguess': 3,'xlock': 3,'xsnoop': 3,'httptunn
el': 3,
'buffer_overflow': 4,'loadmodule': 4,'perl': 4,'r
ootkit': 4,'ps': 4,'sqlattack': 4,'xterm': 4})
newlabeldf_test=labeldf_test.replace({ 'normal' : 0, 'neptune' : 1 ,'back':
1, 'land': 1, 'pod': 1, 'smurf': 1, 'teardrop': 1,'mailbomb': 1, 'apache2':
1, 'processtable': 1, 'udpstorm': 1, 'worm': 1,
'ipsweep' : 2,'nmap' : 2,'portsweep' : 2,'satan'
: 2,'mscan' : 2,'saint' : 2
,'ftp_write': 3,'guess_passwd': 3,'imap': 3,'mult
ihop': 3,'phf': 3,'spy': 3,'warezclient': 3,'warezmaster': 3,'sendmail': 3,'
named': 3,'snmpgetattack': 3,'snmpguess': 3,'xlock': 3,'xsnoop': 3,'httptunn
el': 3,
'buffer_overflow': 4,'loadmodule': 4,'perl': 4,'r
ootkit': 4,'ps': 4,'sqlattack': 4,'xterm': 4})
newdf['label'] = newlabeldf
newdf_test['label'] = newlabeldf_test
print(newdf['label'].head())
to_drop_DoS = [2,3,4]
to_drop_Probe = [1,3,4]
to_drop_R2L = [1,2,4]
to_drop_U2R = [1,2,3]
DoS_df=newdf[~newdf['label'].isin(to_drop_DoS)];
Probe_df=newdf[~newdf['label'].isin(to_drop_Probe)];
R2L_df=newdf[~newdf['label'].isin(to_drop_R2L)];
U2R_df=newdf[~newdf['label'].isin(to_drop_U2R)];
DoS_df_test=newdf_test[~newdf_test['label'].isin(to_drop_DoS)];
Probe_df_test=newdf_test[~newdf_test['label'].isin(to_drop_Probe)];
R2L_df_test=newdf_test[~newdf_test['label'].isin(to_drop_R2L)];
U2R_df_test=newdf_test[~newdf_test['label'].isin(to_drop_U2R)];
print('Train:')
print('Dimensions of DoS:' ,DoS_df.shape)
print('Dimensions of Probe:' ,Probe_df.shape)
print('Dimensions of R2L:' ,R2L_df.shape)
print('Dimensions of U2R:' ,U2R_df.shape)
print('Test:')
print('Dimensions of DoS:' ,DoS_df_test.shape)
print('Dimensions of Probe:' ,Probe_df_test.shape)
print('Dimensions of R2L:' ,R2L_df_test.shape)
print('Dimensions of U2R:' ,U2R_df_test.shape)
X_DoS = DoS_df.drop('label',1)
Y_DoS = DoS_df.label
X_Probe = Probe_df.drop('label',1)
Y_Probe = Probe_df.label
X_R2L = R2L_df.drop('label',1)
Y_R2L = R2L_df.label
X_U2R = U2R_df.drop('label',1)
Y_U2R = U2R_df.label
X_DoS_test = DoS_df_test.drop('label',1)
Y_DoS_test = DoS_df_test.label
X_Probe_test = Probe_df_test.drop('label',1)
Y_Probe_test = Probe_df_test.label
X_R2L_test = R2L_df_test.drop('label',1)
Y_R2L_test = R2L_df_test.label
X_U2R_test = U2R_df_test.drop('label',1)
Y_U2R_test = U2R_df_test.label
colNames=list(X_DoS)
colNames_test=list(X_DoS_test)
from sklearn.feature_selection import SelectPercentile, f_classif
np.seterr(divide='ignore', invalid='ignore');
selector=SelectPercentile(f_classif, percentile=10)
X_newDoS = selector.fit_transform(X_DoS,Y_DoS)
X_newDoS.shape
true=selector.get_support()
newcolindex_DoS=[i for i, x in enumerate(true) if x]
newcolname_DoS=list( colNames[i] for i in newcolindex_DoS )
newcolname_DoS
X_newProbe = selector.fit_transform(X_Probe,Y_Probe)
X_newProbe.shape
true=selector.get_support()
newcolindex_Probe=[i for i, x in enumerate(true) if x]
newcolname_Probe=list( colNames[i] for i in newcolindex_Probe )
newcolname_Probe
X_newR2L = selector.fit_transform(X_R2L,Y_R2L)
X_newR2L.shape
true=selector.get_support()
newcolindex_R2L=[i for i, x in enumerate(true) if x]
newcolname_R2L=list( colNames[i] for i in newcolindex_R2L)
newcolname_R2L
X_newU2R = selector.fit_transform(X_U2R,Y_U2R)
X_newU2R.shape
true=selector.get_support()
newcolindex_U2R=[i for i, x in enumerate(true) if x]
newcolname_U2R=list( colNames[i] for i in newcolindex_U2R)
newcolname_U2R
print('Features selected for DoS:',newcolname_DoS)
print()
print('Features selected for Probe:',newcolname_Probe)
print()
print('Features selected for R2L:',newcolname_R2L)
print()
print('Features selected for U2R:',newcolname_U2R)
from sklearn.feature_selection import RFE
from sklearn.tree import DecisionTreeClassifier
clf = DecisionTreeClassifier(random_state=0)
#rank all features, i.e continue the elimination until the last one
rfe = RFE(clf, n_features_to_select=1)
Y_DoS=Y_DoS.astype('int')
rfe.fit(X_newDoS, Y_DoS)
print ("DoS Features sorted by their rank:")
print (sorted(zip(map(lambda x: round(x, 4), rfe.ranking_), newcolname_DoS))
)
Y_Probe=Y_Probe.astype('int')
rfe.fit(X_newProbe, Y_Probe)
print ("Probe Features sorted by their rank:")
print (sorted(zip(map(lambda x: round(x, 4), rfe.ranking_), newcolname_Probe
)))
Y_R2L=Y_R2L.astype('int')
rfe.fit(X_newR2L, Y_R2L)
print ("R2L Features sorted by their rank:")
print (sorted(zip(map(lambda x: round(x, 4), rfe.ranking_), newcolname_R2L))
)
Y_U2R=Y_U2R.astype('int')
X_newU2R=X_newU2R.astype('int')
rfe.fit(X_newU2R, Y_U2R)
print ("U2R Features sorted by their rank:")
print (sorted(zip(map(lambda x: round(x, 4), rfe.ranking_), newcolname_U2R))
)
from sklearn.feature_selection import RFE
clf = DecisionTreeClassifier(random_state=0)
rfe = RFE(estimator=clf, n_features_to_select=13, step=1)
Y_DoS=Y_DoS.astype('int')
rfe.fit(X_DoS, Y_DoS)
X_rfeDoS=rfe.transform(X_DoS)
true=rfe.support_
rfecolindex_DoS=[i for i, x in enumerate(true) if x]
rfecolname_DoS=list(colNames[i] for i in rfecolindex_DoS)
Y_Probe=Y_Probe.astype('int')
rfe.fit(X_Probe, Y_Probe)
X_rfeProbe=rfe.transform(X_Probe)
true=rfe.support_
rfecolindex_Probe=[i for i, x in enumerate(true) if x]
rfecolname_Probe=list(colNames[i] for i in rfecolindex_Probe)
Y_R2L=Y_R2L.astype('int')
rfe.fit(X_R2L, Y_R2L)
X_rfeR2L=rfe.transform(X_R2L)
true=rfe.support_
rfecolindex_R2L=[i for i, x in enumerate(true) if x]
rfecolname_R2L=list(colNames[i] for i in rfecolindex_R2L)
Y_U2R=Y_U2R.astype('int')
rfe.fit(X_U2R, Y_U2R)
X_rfeU2R=rfe.transform(X_U2R)
true=rfe.support_
rfecolindex_U2R=[i for i, x in enumerate(true) if x]
rfecolname_U2R=list(colNames[i] for i in rfecolindex_U2R)
print('Features selected for DoS:',rfecolname_DoS)
print()
print('Features selected for Probe:',rfecolname_Probe)
print()
print('Features selected for R2L:',rfecolname_R2L)
print()
print('Features selected for U2R:',rfecolname_U2R)
print(X_rfeDoS.shape)
print(X_rfeProbe.shape)
print(X_rfeR2L.shape)
print(X_rfeU2R.shape)
# all features
clf_DoS=DecisionTreeClassifier(random_state=0)
clf_Probe=DecisionTreeClassifier(random_state=0)
clf_R2L=DecisionTreeClassifier(random_state=0)
clf_U2R=DecisionTreeClassifier(random_state=0)
clf_DoS.fit(X_DoS, Y_DoS)
clf_Probe.fit(X_Probe, Y_Probe)
clf_R2L.fit(X_R2L, Y_R2L)
clf_U2R.fit(X_U2R, Y_U2R)
# selected features
clf_rfeDoS=DecisionTreeClassifier(random_state=0)
clf_rfeProbe=DecisionTreeClassifier(random_state=0)
clf_rfeR2L=DecisionTreeClassifier(random_state=0)
clf_rfeU2R=DecisionTreeClassifier(random_state=0)
clf_rfeDoS.fit(X_rfeDoS, Y_DoS)
clf_rfeProbe.fit(X_rfeProbe, Y_Probe)
clf_rfeR2L.fit(X_rfeR2L, Y_R2L)
clf_rfeU2R.fit(X_rfeU2R, Y_U2R)
# reduce test dataset to 13 features, use only features described in rfecoln
ame_DoS etc.
X_DoS_test2=X_DoS_test[:,rfecolindex_DoS]
X_Probe_test2=X_Probe_test[:,rfecolindex_Probe]
X_R2L_test2=X_R2L_test[:,rfecolindex_R2L]
X_U2R_test2=X_U2R_test[:,rfecolindex_U2R]
X_U2R_test2.shape
Y_DoS_pred2=clf_rfeDoS.predict(X_DoS_test2)
# Create confusion matrix
pd.crosstab(Y_DoS_test, Y_DoS_pred2, rownames=['Actual attacks'], colnames=[
'Predicted attacks'])
Y_Probe_pred2=clf_rfeProbe.predict(X_Probe_test2)
# Create confusion matrix
pd.crosstab(Y_Probe_test, Y_Probe_pred2, rownames=['Actual attacks'], colnam
es=['Predicted attacks'])
Y_R2L_pred2=clf_rfeR2L.predict(X_R2L_test2)
# Create confusion matrix
pd.crosstab(Y_R2L_test, Y_R2L_pred2, rownames=['Actual attacks'], colnames=[
'Predicted attacks'])
Y_U2R_pred2=clf_rfeU2R.predict(X_U2R_test2)
# Create confusion matrix
pd.crosstab(Y_U2R_test, Y_U2R_pred2, rownames=['Actual attacks'], colnames=[
'Predicted attacks'])