本文共 8757 字，大约阅读时间需要 29 分钟。

用于做网络流量分类

import tensorflow as tfimport numpy as npimport osfrom tensorflow.keras.models import Sequentialfrom tensorflow.keras import layersfrom sklearn import metrics, neighborsfrom tensorflow import kerasfrom sklearn.svm import SVCfrom sklearn.naive_bayes import GaussianNBfrom sklearn.neighbors import KNeighborsClassifierfrom sklearn.tree import DecisionTreeClassifierfrom sklearn.metrics import confusion_matrix, roc_curveimport matplotlib.pyplot as pltfrom sklearn.model_selection import train_test_splitimport timefrom sklearn.ensemble import RandomForestClassifier

数据处理

list_y = ['WWW','MAIL','FTP-CONTROL','FTP-PASV','ATTACK','P2P','DATABASE','FTP-DATA','MULTIMEDIA','SERVICES','INTERACTIVE','GAMES']#read the file,change 'Y,N,?,', translate to tensordef data_prepross(filename):    X, Y = [], []    for f in filename:        print(f)        with open(os.getcwd() + '/' + f, 'r') as file:            for n, i in enumerate(file.readlines()[253:]):                i = i.replace('Y','1')                i = i.replace('N', '0')                spl = i.split(',')                if spl.count('?')>8:                    continue                i = i.replace('\n', '')                fz = [float(f) for f in i.split(',')[:-1] if f != '?']                meana = sum(fz) / len(fz)                i = i.replace('?', str(0))                #均值填充，加高斯白噪声                x = [float(j) for j in i.split(',')[:-1]] +[meana]*8  +  np.random.normal(0,1,256)                #x = [float(j) for j in i.split(',')[:-1]] + [0] * 8                #x =x.tolist()                y = i.split(',')[-1].replace('FTP-CO0TROL','FTP-CONTROL')                y = y.replace('I0TERACTIVE','INTERACTIVE' )                y = list_y.index(y)                X.append(x)                Y.append(y)            file.close()    return X, Y#data nomalization#train_x,train_y = data_prepross(['entry01.weka.allclass.arff',])total_x,total_y = data_prepross(['entry01.weka.allclass.arff','entry02.weka.allclass.arff','entry03.weka.allclass.arff','entry04.weka.allclass.arff',                                 'entry05.weka.allclass.arff','entry09.weka.allclass.arff', 'entry10.weka.allclass.arff','entry07.weka.allclass.arff','entry08.weka.allclass.arff','entry06.weka.allclass.arff'])train_x,test_x,train_y,test_y = train_test_split(total_x,total_y,test_size=0.25, random_state=0)train_x = tf.convert_to_tensor(train_x, dtype=tf.float32)train_y= tf.convert_to_tensor(train_y,dtype=tf.int32)test_x = tf.convert_to_tensor(test_x, dtype=tf.float32)test_y = tf.convert_to_tensor(test_y,dtype= tf.int32)train_x = tf.keras.utils.normalize(train_x, axis=1)test_x = tf.keras.utils.normalize(test_x, axis=1)

BP神经网络

num_classes =12num_pixels=256def baseline_model():     model = Sequential()     model.add(layers.Dense(num_pixels, input_dim=num_pixels,  activation='relu'))     #layers.Dropout(0.5)     model.add(layers.Dense(num_classes,  activation='softmax'))     model.summary()     model.compile(loss='sparse_categorical_crossentropy', optimizer='adam', metrics=['accuracy'])     return model#,class_weight=class_weightdef baseline():    t1 = time.time()    model= baseline_model()    X_train = tf.reshape(train_x,[-1,256])    X_test =  tf.reshape(test_x, [-1,256])    history = model.fit(X_train, train_y, validation_data=(X_test, test_y), nb_epoch=20, batch_size=batchsize, verbose=2,)    scores = model.evaluate(X_test, test_y, verbose=0)    predict_y = model.predict(X_test)    t2 = time.time()    print("Baseline Error: %.2f%%" % (100-scores[1]*100),t2-t1)    print(history.history)    return scores,predict_y

#CNN

def simple_CNNmodel():    model = keras.models.Sequential([        layers.Conv2D(filters=8, kernel_size=(3, 3),padding='same',input_shape=(16, 16, 1), activation='relu'),        layers.MaxPooling2D(pool_size=(2, 2), padding = 'same' ),        layers.Conv2D(filters=16, kernel_size=(3, 3), padding='same',activation='relu'),        layers.MaxPooling2D(pool_size=(2, 2), padding = 'same' ),        #layers.Dropout(0.25),        #(5,5,16) > 400        layers.Flatten(),        layers.Dense(256, activation='relu'),        #layers.Dropout(0.5),        #layers.Dense(84, activation='relu'),        layers.Dense(128, activation='relu'),        #layers.Dropout(0.5),        layers.Dense(12, activation='softmax')    ])# Compile model    model.compile(loss="sparse_categorical_crossentropy", optimizer='adam', metrics=['accuracy'])    return model#,class_weight=class_weightdef simple_CNN():    t1 = time.time()    model = simple_CNNmodel()    X_train = tf.reshape(train_x,[-1,16,16,1])    X_test =  tf.reshape(test_x, [-1,16,16,1])    model.summary()    history = model.fit(X_train, train_y, validation_data=(X_test, test_y), nb_epoch=25, batch_size=128, verbose=2)    scores = model.evaluate(X_test, test_y, verbose=0)    t2 = time.time()    pred_y = model.predict(X_test)    print(scores)    print("Baseline Error: %.2f%%" % (100 - scores[1] * 100),t2-t1)    print(history.history)    return scores,pred_y

朴素贝叶斯

def Bayes(trainData,trainLable,testData,testLable):    t1 = time.time()    mnb = GaussianNB()  #    mnb.fit(trainData, trainLable)  #    y_predict = mnb.predict(testData)    t2 =time.time()    print(t2-t1)    print(confusion_matrix(testLable,y_predict))    print('The Accuracy of Naive Bayes Classifier is:', mnb.score(testData,testLable ))

决策树

def DecisionTr(trainData,trainLable,testData,testLable):    t1 = time.time()    model = DecisionTreeClassifier()    model.fit(trainData, trainLable)    predicted = model.predict(testData)    score = metrics.accuracy_score(testLable, predicted)    t2 = time.time()    print(t2-t1,score)    print(confusion_matrix(testLable,predicted))

SVM支持向量机

def SVM(trainData,trainLable,testData,testLable):    t1=time.time()    clf = SVC()    clf.fit(trainData, trainLable)    svmPredict=clf.predict(testData)    svmScore=metrics.accuracy_score(testLable, svmPredict)    t2=time.time()    print(t2-t1,svmScore)

KNN

def Knn(trainData,trainLable,testData,testLable):    t1=time.time()    knn = KNeighborsClassifier()    knn.fit(trainData, trainLable)    knnPredict = knn.predict(testData)    knnscore=metrics.accuracy_score(testLable, knnPredict)    t2=time.time()    print(t2-t1,knnscore)    print(confusion_matrix(testLable, knnPredict))Knn(train_x.numpy(),train_y.numpy(),test_x.numpy(),test_y.numpy())

灰度图片绘图

def plt_image(trainx,trainy):    p_www = np.where(trainy == 0)[0][0]    p_mail = np.where(trainy == 1)[0][0]    p_control = np.where(trainy == 2)[0][0]    p_pasv = np.where(trainy == 3)[0][0]    p_attack = np.where(trainy == 4)[0][0]    p_p2p = np.where(trainy == 5)[0][0]    p_database = np.where(trainy == 6)[0][0]    p_data = np.where(trainy == 7)[0][0]    p_multimedia = np.where(trainy == 8)[0][0]    p_service = np.where(trainy == 9)[0][0]    p_interactive = np.where(trainy == 10)[0][0]    p_games = np.where(trainy == 11)[0][0]    plt.figure(num='classffication', figsize=(6, 12))    plt.subplot(3,4,1)    plt.title('WWW')    plt.imshow(np.reshape(trainx[p_www],(16,16)))    plt.subplot(3,4,2)    plt.title('MAIL')    plt.imshow(np.reshape(trainx[p_mail],(16,16)))    plt.subplot(3,4,3)    plt.title('FTP-CONTROL')    plt.imshow(np.reshape(trainx[p_control],(16,16)))    plt.subplot(3,4,4)    plt.title('FTP-PASV')    plt.imshow(np.reshape(trainx[p_pasv],(16,16)))    plt.subplot(3,4,5)    plt.title('ATTCK')    plt.imshow(np.reshape(trainx[p_attack],(16,16)))    plt.subplot(3,4,6)    plt.title('P2P')    plt.imshow(np.reshape(trainx[p_p2p],(16,16)))    plt.subplot(3,4,7)    plt.title('DATABASE')    plt.imshow(np.reshape(train_x[p_database],(16,16)))    plt.subplot(3,4,8)    plt.title('FTP-DATA')    plt.imshow(np.reshape(trainx[p_data],(16,16)))    plt.subplot(3,4,9)    plt.title('MULTIMEDIA')    plt.imshow(np.reshape(train_x[p_multimedia],(16,16)))    plt.subplot(3,4,10)    plt.title('SERVICES')    plt.imshow(np.reshape(trainx[p_service],(16,16)))    plt.subplot(3,4,11)    plt.title('INTERACTIVE')    plt.imshow(np.reshape(train_x[p_interactive],(16,16)))    plt.subplot(3,4,12)    plt.title('GAMES')    plt.imshow(np.reshape(trainx[p_games],(16,16)))    plt.show()

混淆矩阵绘图

`def plot_confusion_matrix(title, pred_y):    cm = confusion_matrix(test_y,np.argmax(pred_y, 1))    #cm = confusion_matrix(test_y, pred_y> 0.5)    labels_name = list_y    cm = cm.astype('float') / cm.sum(axis=1)[:, np.newaxis]    # 归一化    plt.imshow(cm, interpolation='nearest')    # 在特定的窗口上显示图像    plt.title(title)    # 图像标题    plt.colorbar()    num_local = np.array(range(len(labels_name)))    plt.xticks(num_local, labels_name, rotation=90)    # 将标签印在x轴坐标上    plt.yticks(num_local, labels_name)    # 将标签印在y轴坐标上    plt.ylabel('True')    plt.xlabel('Predicted')    plt.show()

转载地址：http://raben.baihongyu.com/