基于网络安全问题的洋葱路由流量识别

#安装相关的第三方包!pip install scikit-learn
!pip install pandas
!pip install cicflowmeter

<br/>

!sudo apt-get update
!sudo apt-get install wireshark

!sudo wireshark

#!cicflowmeter -f *.pcap -c *.csv #通过该命令将pcap格式的流量抓包文件转化为csv格式的特征文件（*号为文件名）。

import pandas as pdfrom sklearn.preprocessing import StandardScalerfrom sklearn import linear_modelimport numpy as npfrom sklearn.utils import shufflefrom sklearn import svmfrom sklearn import preprocessing

file = pd.read_csv('data/data150018/SelectedFeatures-10s-TOR-NonTOR.csv',skipinitialspace=True)print(file.columns)with pd.option_context('mode.use_inf_as_na', True):
    file = file.dropna()#file = shuffle(file)y = file.iloc[:,-1]
standard_scaler = StandardScaler()

allX = file.drop(['Source IP','Destination IP','label'],axis=1)
column = allX.columns

from sklearn.model_selection import train_test_split
x_train, x_test, y_train, y_test = train_test_split(allX, y,test_size=0.25)

from sklearn.decomposition import PCA
feature_reducted_train_data = PCA(n_components=24).fit_transform(x_train)
feature_reducted_test_data = PCA(n_components=24).fit_transform(x_test)

from sklearn.ensemble import AdaBoostClassifierfrom sklearn.neighbors import KNeighborsClassifierfrom sklearn import svm#使用KNN算法进行训练和预测knn = KNeighborsClassifier(n_neighbors=2)
knn.fit(feature_reducted_train_data, y_train)
knn_predict = knn.predict(feature_reducted_test_data)#使用Adaboost算法进行训练和预测randomforest = AdaBoostClassifier()
randomforest.fit(feature_reducted_train_data, y_train)
randomforest_predict = randomforest.predict(feature_reducted_test_data)#使用SVM算法进行训练和预测svm_object = svm.SVC(kernel='linear', max_iter=1000000,C=0.1)
svm_object.fit(feature_reducted_train_data, y_train)
svm_predict = svm_object.predict(feature_reducted_test_data)

from sklearn import metricsprint(metrics.classification_report(y_test, knn_predict))print(metrics.classification_report(y_test, randomforest_predict))print(metrics.classification_report(y_test, svm_predict))