3.2 OCGAN setting

# import modules
%matplotlib inline
import os
import random
import numpy as np
import pandas as pd
import matplotlib.pyplot as plt
import seaborn as sns
from tqdm import tqdm_notebook as tqdm
from keras.models import Model
from keras.layers import Input, Reshape
from keras.layers.core import Dense, Activation, Dropout, Flatten
from keras.layers.normalization import BatchNormalization
from keras.layers.convolutional import UpSampling1D, Conv1D
from keras.layers.advanced_activations import LeakyReLU
from keras.optimizers import Adam, SGD
from keras.callbacks import TensorBoard
from sklearn.preprocessing import StandardScaler

# set parameters
dim = f_x.shape[1]
num = f_x.shape[0]
g_data = f_x

# Standard Scaler
ss = StandardScaler()
g_data = pd.DataFrame(ss.fit_transform(g_data))

# generator
def get_generative(G_in, dense_dim=200, out_dim= dim, lr=1e-3):
    x = Dense(dense_dim)(G_in)
    x = Activation('tanh')(x)
    G_out = Dense(out_dim, activation='tanh')(x)
    G = Model(G_in, G_out)
    opt = SGD(lr=lr)
    G.compile(loss='binary_crossentropy', optimizer=opt)
    return G, G_out

G_in = Input(shape=[10])
G, G_out = get_generative(G_in)
G.summary()

# discriminator
def get_discriminative(D_in, lr=1e-3, drate=.25, n_channels= dim, conv_sz=5, leak=.2):
    x = Reshape((-1, 1))(D_in)
    x = Conv1D(n_channels, conv_sz, activation='relu')(x)
    x = Dropout(drate)(x)
    x = Flatten()(x)
    x = Dense(n_channels)(x)
    D_out = Dense(2, activation='sigmoid')(x)
    D = Model(D_in, D_out)
    dopt = Adam(lr=lr)
    D.compile(loss='binary_crossentropy', optimizer=dopt)
    return D, D_out

D_in = Input(shape=[dim])
D, D_out = get_discriminative(D_in)
D.summary()

# set up gan
def set_trainability(model, trainable=False):
    model.trainable = trainable
    for layer in model.layers:
        layer.trainable = trainable
        
def make_gan(GAN_in, G, D):
    set_trainability(D, False)
    x = G(GAN_in)
    GAN_out = D(x)
    GAN = Model(GAN_in, GAN_out)
    GAN.compile(loss='binary_crossentropy', optimizer=G.optimizer)
    return GAN, GAN_out

GAN_in = Input([10])
GAN, GAN_out = make_gan(GAN_in, G, D)
GAN.summary()

# pre train
def sample_data_and_gen(G, noise_dim=10, n_samples= num):
    XT = np.array(g_data)
    XN_noise = np.random.uniform(0, 1, size=[n_samples, noise_dim])
    XN = G.predict(XN_noise)
    X = np.concatenate((XT, XN))
    y = np.zeros((2*n_samples, 2))
    y[:n_samples, 1] = 1
    y[n_samples:, 0] = 1
    return X, y

def pretrain(G, D, noise_dim=10, n_samples = num, batch_size=32):
    X, y = sample_data_and_gen(G, n_samples=n_samples, noise_dim=noise_dim)
    set_trainability(D, True)
    D.fit(X, y, epochs=1, batch_size=batch_size)
    
pretrain(G, D)

def sample_noise(G, noise_dim=10, n_samples=num):
    X = np.random.uniform(0, 1, size=[n_samples, noise_dim])
    y = np.zeros((n_samples, 2))
    y[:, 1] = 1
    return X, y

# one class detector
def oneclass(data,kernel = 'rbf',gamma = 'auto'):
    num1 = int(len(data)/2)
    num2 = int(len(data)+1)
    from sklearn import svm
    clf = svm.OneClassSVM(kernel=kernel, gamma=gamma).fit(data[0:num1])
    origin = pd.DataFrame(clf.score_samples(data[0:num1]))
    new = pd.DataFrame(clf.score_samples(data[num1:num2]))

    occ = pd.concat([pd.DataFrame(new[0] < origin[0].min()),pd.DataFrame(new[0] > origin[0].max())], axis=1)
    occ['ava'] = pd.DataFrame(occ.iloc[:,1:2] == occ.iloc[:,0:1])
    err = sum(occ['ava'] == False)/len(occ['ava'])
    return err

# training
def train1(GAN, G, D, epochs=1, n_samples= num, noise_dim=10, batch_size=32, verbose=False, v_freq=dim,):
    d_loss = []
    g_loss = []
    e_range = range(epochs)
    if verbose:
        e_range = tqdm(e_range)
    for epoch in e_range:
        X, y = sample_data_and_gen(G, n_samples=n_samples, noise_dim=noise_dim)
        set_trainability(D, True)
        d_loss.append(D.train_on_batch(X, y))
        xx,yy = X,y
        err = oneclass(xx) 
        print("The %d times epoch one class svm Error Rate=%f" %(epoch, err))
        
        X, y = sample_noise(G, n_samples=n_samples, noise_dim=noise_dim)
        set_trainability(D, False)
        g_loss.append(GAN.train_on_batch(X, y))
        if verbose and (epoch + 1) % v_freq == 0:
            print("Epoch #{}: Generative Loss: {}, Discriminative Loss: {}".format(epoch + 1, g_loss[-1], d_loss[-1]))
    return d_loss, g_loss, xx, yy

WARNING:tensorflow:From C:3-gpu-packagesimpl.py:180: add_dispatch_support..wrapper (from tensorflow.python.ops.array_ops) is deprecated and will be removed in a future version. Instructions for updating: Use tf.where in 2.0, which has the same broadcast rule as np.where ________________________________________________________________ Layer (type) Output Shape Param #
================================================================= input_4 (InputLayer) (None, 10) 0
_________________________________________________________________ dense_5 (Dense) (None, 200) 2200
_________________________________________________________________ activation_2 (Activation) (None, 200) 0
_________________________________________________________________ dense_6 (Dense) (None, 66) 13266
================================================================= Total params: 15,466 Trainable params: 15,466 Non-trainable params: 0 _________________________________________________________________ _________________________________________________________________ Layer (type) Output Shape Param #
================================================================= input_5 (InputLayer) (None, 66) 0
_________________________________________________________________ reshape_2 (Reshape) (None, 66, 1) 0
_________________________________________________________________ conv1d_2 (Conv1D) (None, 62, 66) 396
_________________________________________________________________ dropout_2 (Dropout) (None, 62, 66) 0
_________________________________________________________________ flatten_2 (Flatten) (None, 4092) 0
_________________________________________________________________ dense_7 (Dense) (None, 66) 270138
_________________________________________________________________ dense_8 (Dense) (None, 2) 134
================================================================= Total params: 270,668 Trainable params: 270,668 Non-trainable params: 0 _________________________________________________________________ _________________________________________________________________ Layer (type) Output Shape Param #
================================================================= input_6 (InputLayer) (None, 10) 0
_________________________________________________________________ model_4 (Model) (None, 66) 15466
_________________________________________________________________ model_5 (Model) (None, 2) 270668
================================================================= Total params: 286,134 Trainable params: 15,466 Non-trainable params: 270,668 _________________________________________________________________ Epoch 1/1 92/92 [==============================] - ETA: 1s - loss: 0.667 - 1s 6ms/step - loss: 0.5688

3.3 SMOTE setting

因正負樣本相差過小，且smote function為自動平衡方法，也就是說它會自動抓取正負樣本差距給予平衡，但因差距僅三樣，因此我將資料0的部分重複貼上依次，使其差距擴大至49+3 = 52，下一步驟的adaysn也採用相同的更動資料。

x0 = x[y['CRE'].isin([0])]
y0 = y[y['CRE'].isin([0])]
x0 = pd.concat([x,x0],axis = 0)  
y0 = pd.concat([y,y0],axis = 0)  

from imblearn.over_sampling import SMOTE
smo = SMOTE()
X_smo, y_smo = smo.fit_sample(x0, y0)
X_smo = pd.DataFrame(X_smo)
y_smo = pd.DataFrame(y_smo)

C:3-gpu-packages.py:724: DataConversionWarning: A column-vector y was passed when a 1d array was expected. Please change the shape of y to (n_samples, ), for example using ravel(). y = column_or_1d(y, warn=True)

%matplotlib inline
by_fraud = y_smo.groupby(0)
by_fraud.size().plot(kind = 'bar')

<matplotlib.axes._subplots.AxesSubplot at 0x1a7c68d49e8>

3. 4 ADASYN setting

from imblearn.over_sampling import ADASYN
ada = ADASYN()
X_ada, y_ada = ada.fit_sample(x0, y0)
X_ada = pd.DataFrame(X_ada)
y_ada = pd.DataFrame(y_ada)

C:3-gpu-packages.py:724: DataConversionWarning: A column-vector y was passed when a 1d array was expected. Please change the shape of y to (n_samples, ), for example using ravel(). y = column_or_1d(y, warn=True)

ocwgan epoch 1

d_loss, g_loss ,xx,yy= train(GAN, G, D, epochs=1, verbose=True)
new_data = gen(GAN, G, D, times = 100,verbose=True)
test_y_predicted = clf.predict(new_data)
print(test_y_predicted,'acc = ',np.mean(test_y_predicted))

HBox(children=(IntProgress(value=0, max=1), HTML(value=’’)))

The 0 times epoch one class svm Error Rate=1.000000

The 0 times generator one class svm Error Rate=1.000000 The 1 times generator one class svm Error Rate=1.000000 The 2 times generator one class svm Error Rate=1.000000 The 3 times generator one class svm Error Rate=1.000000 The 4 times generator one class svm Error Rate=1.000000 The 5 times generator one class svm Error Rate=1.000000 The 6 times generator one class svm Error Rate=1.000000 The 7 times generator one class svm Error Rate=1.000000 The 8 times generator one class svm Error Rate=1.000000 The 9 times generator one class svm Error Rate=1.000000 The 10 times generator one class svm Error Rate=1.000000 The 11 times generator one class svm Error Rate=1.000000 The 12 times generator one class svm Error Rate=0.978261 The 13 times generator one class svm Error Rate=1.000000 The 14 times generator one class svm Error Rate=1.000000 The 15 times generator one class svm Error Rate=0.978261 The 16 times generator one class svm Error Rate=1.000000 The 17 times generator one class svm Error Rate=1.000000 The 18 times generator one class svm Error Rate=1.000000 The 19 times generator one class svm Error Rate=0.978261 The 20 times generator one class svm Error Rate=1.000000 The 21 times generator one class svm Error Rate=1.000000 The 22 times generator one class svm Error Rate=1.000000 The 23 times generator one class svm Error Rate=1.000000 The 24 times generator one class svm Error Rate=1.000000 The 25 times generator one class svm Error Rate=0.978261 The 26 times generator one class svm Error Rate=1.000000 The 27 times generator one class svm Error Rate=1.000000 The 28 times generator one class svm Error Rate=1.000000 The 29 times generator one class svm Error Rate=0.978261 The 30 times generator one class svm Error Rate=1.000000 The 31 times generator one class svm Error Rate=0.978261 The 32 times generator one class svm Error Rate=1.000000 The 33 times generator one class svm Error Rate=1.000000 The 34 times generator one class svm Error Rate=1.000000 The 35 times generator one class svm Error Rate=1.000000 The 36 times generator one class svm Error Rate=1.000000 The 37 times generator one class svm Error Rate=1.000000 The 38 times generator one class svm Error Rate=1.000000 The 39 times generator one class svm Error Rate=0.978261 The 40 times generator one class svm Error Rate=1.000000 The 41 times generator one class svm Error Rate=1.000000 The 42 times generator one class svm Error Rate=1.000000 The 43 times generator one class svm Error Rate=1.000000 The 44 times generator one class svm Error Rate=1.000000 The 45 times generator one class svm Error Rate=1.000000 The 46 times generator one class svm Error Rate=1.000000 The 47 times generator one class svm Error Rate=1.000000 The 48 times generator one class svm Error Rate=0.978261 The 49 times generator one class svm Error Rate=1.000000 The 50 times generator one class svm Error Rate=1.000000 The 51 times generator one class svm Error Rate=1.000000 The 52 times generator one class svm Error Rate=1.000000 The 53 times generator one class svm Error Rate=1.000000 The 54 times generator one class svm Error Rate=1.000000 The 55 times generator one class svm Error Rate=1.000000 The 56 times generator one class svm Error Rate=1.000000 The 57 times generator one class svm Error Rate=1.000000 The 58 times generator one class svm Error Rate=0.978261 The 59 times generator one class svm Error Rate=1.000000 The 60 times generator one class svm Error Rate=0.956522 The 61 times generator one class svm Error Rate=1.000000 The 62 times generator one class svm Error Rate=1.000000 The 63 times generator one class svm Error Rate=1.000000 The 64 times generator one class svm Error Rate=1.000000 The 65 times generator one class svm Error Rate=1.000000 The 66 times generator one class svm Error Rate=1.000000 The 67 times generator one class svm Error Rate=1.000000 The 68 times generator one class svm Error Rate=1.000000 The 69 times generator one class svm Error Rate=1.000000 The 70 times generator one class svm Error Rate=1.000000 The 71 times generator one class svm Error Rate=1.000000 The 72 times generator one class svm Error Rate=1.000000 The 73 times generator one class svm Error Rate=1.000000 The 74 times generator one class svm Error Rate=1.000000 The 75 times generator one class svm Error Rate=0.978261 The 76 times generator one class svm Error Rate=1.000000 The 77 times generator one class svm Error Rate=1.000000 The 78 times generator one class svm Error Rate=1.000000 The 79 times generator one class svm Error Rate=1.000000 The 80 times generator one class svm Error Rate=1.000000 The 81 times generator one class svm Error Rate=1.000000 The 82 times generator one class svm Error Rate=1.000000 The 83 times generator one class svm Error Rate=1.000000 The 84 times generator one class svm Error Rate=1.000000 The 85 times generator one class svm Error Rate=0.978261 The 86 times generator one class svm Error Rate=1.000000 The 87 times generator one class svm Error Rate=1.000000 The 88 times generator one class svm Error Rate=1.000000 The 89 times generator one class svm Error Rate=1.000000 The 90 times generator one class svm Error Rate=1.000000 The 91 times generator one class svm Error Rate=1.000000 The 92 times generator one class svm Error Rate=1.000000 The 93 times generator one class svm Error Rate=0.978261 The 94 times generator one class svm Error Rate=1.000000 The 95 times generator one class svm Error Rate=1.000000 The 96 times generator one class svm Error Rate=0.978261 The 97 times generator one class svm Error Rate=1.000000 The 98 times generator one class svm Error Rate=1.000000 The 99 times generator one class svm Error Rate=0.978261 [1 1 1 … 1 1 1] acc = 1.0

ocwgan epoch 2

d_loss, g_loss ,xx,yy= train(GAN, G, D, epochs=5, verbose=True)
new_data = gen(GAN, G, D, times = 100,verbose=True)
test_y_predicted = clf.predict(new_data)
test_y_predicted
print(test_y_predicted,'acc = ',np.mean(test_y_predicted))

HBox(children=(IntProgress(value=0, max=5), HTML(value=’’)))

The 0 times epoch one class svm Error Rate=1.000000 The 1 times epoch one class svm Error Rate=0.673913 The 2 times epoch one class svm Error Rate=0.217391 The 3 times epoch one class svm Error Rate=0.000000 The 4 times epoch one class svm Error Rate=0.021739

The 0 times generator one class svm Error Rate=0.000000 The 1 times generator one class svm Error Rate=0.021739 The 2 times generator one class svm Error Rate=0.000000 The 3 times generator one class svm Error Rate=0.021739 The 4 times generator one class svm Error Rate=0.021739 The 5 times generator one class svm Error Rate=0.000000 The 6 times generator one class svm Error Rate=0.000000 The 7 times generator one class svm Error Rate=0.000000 The 8 times generator one class svm Error Rate=0.021739 The 9 times generator one class svm Error Rate=0.021739 The 10 times generator one class svm Error Rate=0.000000 The 11 times generator one class svm Error Rate=0.000000 The 12 times generator one class svm Error Rate=0.000000 The 13 times generator one class svm Error Rate=0.000000 The 14 times generator one class svm Error Rate=0.000000 The 15 times generator one class svm Error Rate=0.000000 The 16 times generator one class svm Error Rate=0.021739 The 17 times generator one class svm Error Rate=0.000000 The 18 times generator one class svm Error Rate=0.000000 The 19 times generator one class svm Error Rate=0.000000 The 20 times generator one class svm Error Rate=0.000000 The 21 times generator one class svm Error Rate=0.000000 The 22 times generator one class svm Error Rate=0.000000 The 23 times generator one class svm Error Rate=0.000000 The 24 times generator one class svm Error Rate=0.000000 The 25 times generator one class svm Error Rate=0.000000 The 26 times generator one class svm Error Rate=0.000000 The 27 times generator one class svm Error Rate=0.000000 The 28 times generator one class svm Error Rate=0.021739 The 29 times generator one class svm Error Rate=0.000000 The 30 times generator one class svm Error Rate=0.021739 The 31 times generator one class svm Error Rate=0.000000 The 32 times generator one class svm Error Rate=0.000000 The 33 times generator one class svm Error Rate=0.021739 The 34 times generator one class svm Error Rate=0.000000 The 35 times generator one class svm Error Rate=0.000000 The 36 times generator one class svm Error Rate=0.000000 The 37 times generator one class svm Error Rate=0.000000 The 38 times generator one class svm Error Rate=0.000000 The 39 times generator one class svm Error Rate=0.000000 The 40 times generator one class svm Error Rate=0.000000 The 41 times generator one class svm Error Rate=0.000000 The 42 times generator one class svm Error Rate=0.000000 The 43 times generator one class svm Error Rate=0.000000 The 44 times generator one class svm Error Rate=0.000000 The 45 times generator one class svm Error Rate=0.000000 The 46 times generator one class svm Error Rate=0.000000 The 47 times generator one class svm Error Rate=0.000000 The 48 times generator one class svm Error Rate=0.000000 The 49 times generator one class svm Error Rate=0.000000 The 50 times generator one class svm Error Rate=0.000000 The 51 times generator one class svm Error Rate=0.000000 The 52 times generator one class svm Error Rate=0.000000 The 53 times generator one class svm Error Rate=0.000000 The 54 times generator one class svm Error Rate=0.000000 The 55 times generator one class svm Error Rate=0.021739 The 56 times generator one class svm Error Rate=0.000000 The 57 times generator one class svm Error Rate=0.021739 The 58 times generator one class svm Error Rate=0.000000 The 59 times generator one class svm Error Rate=0.000000 The 60 times generator one class svm Error Rate=0.000000 The 61 times generator one class svm Error Rate=0.000000 The 62 times generator one class svm Error Rate=0.000000 The 63 times generator one class svm Error Rate=0.000000 The 64 times generator one class svm Error Rate=0.000000 The 65 times generator one class svm Error Rate=0.000000 The 66 times generator one class svm Error Rate=0.000000 The 67 times generator one class svm Error Rate=0.000000 The 68 times generator one class svm Error Rate=0.000000 The 69 times generator one class svm Error Rate=0.021739 The 70 times generator one class svm Error Rate=0.000000 The 71 times generator one class svm Error Rate=0.000000 The 72 times generator one class svm Error Rate=0.000000 The 73 times generator one class svm Error Rate=0.000000 The 74 times generator one class svm Error Rate=0.000000 The 75 times generator one class svm Error Rate=0.000000 The 76 times generator one class svm Error Rate=0.000000 The 77 times generator one class svm Error Rate=0.000000 The 78 times generator one class svm Error Rate=0.000000 The 79 times generator one class svm Error Rate=0.000000 The 80 times generator one class svm Error Rate=0.000000 The 81 times generator one class svm Error Rate=0.000000 The 82 times generator one class svm Error Rate=0.000000 The 83 times generator one class svm Error Rate=0.000000 The 84 times generator one class svm Error Rate=0.000000 The 85 times generator one class svm Error Rate=0.000000 The 86 times generator one class svm Error Rate=0.000000 The 87 times generator one class svm Error Rate=0.000000 The 88 times generator one class svm Error Rate=0.000000 The 89 times generator one class svm Error Rate=0.000000 The 90 times generator one class svm Error Rate=0.000000 The 91 times generator one class svm Error Rate=0.000000 The 92 times generator one class svm Error Rate=0.000000 The 93 times generator one class svm Error Rate=0.000000 The 94 times generator one class svm Error Rate=0.000000 The 95 times generator one class svm Error Rate=0.021739 The 96 times generator one class svm Error Rate=0.000000 The 97 times generator one class svm Error Rate=0.000000 The 98 times generator one class svm Error Rate=0.000000 The 99 times generator one class svm Error Rate=0.000000 [1 1 1 … 1 1 1] acc = 1.0

ocwgan epoch 3

d_loss, g_loss ,xx,yy= train(GAN, G, D, epochs=3, verbose=True)
new_data = gen(GAN, G, D, times = 1,verbose=True)
test_y_predicted = clf.predict(new_data)
print(test_y_predicted,'acc = ',np.mean(test_y_predicted))

HBox(children=(IntProgress(value=0, max=3), HTML(value=’’)))

The 0 times epoch one class svm Error Rate=0.021739 The 1 times epoch one class svm Error Rate=0.000000 The 2 times epoch one class svm Error Rate=0.000000

The 0 times generator one class svm Error Rate=0.000000 [1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1] acc = 1.0

ocgan epoch 1

d_loss, g_loss ,xx,yy= train1(GAN, G, D, epochs=1, verbose=True)
new_data = gen(GAN, G, D, times = 10,verbose=True)
test_y_predicted = clf.predict(new_data)
print(test_y_predicted,'acc = ',np.mean(test_y_predicted))

HBox(children=(IntProgress(value=0, max=1), HTML(value=’’)))

The 0 times epoch one class svm Error Rate=1.000000

The 0 times generator one class svm Error Rate=1.000000 The 1 times generator one class svm Error Rate=1.000000 The 2 times generator one class svm Error Rate=1.000000 The 3 times generator one class svm Error Rate=1.000000 The 4 times generator one class svm Error Rate=1.000000 The 5 times generator one class svm Error Rate=1.000000 The 6 times generator one class svm Error Rate=1.000000 The 7 times generator one class svm Error Rate=1.000000 The 8 times generator one class svm Error Rate=1.000000 The 9 times generator one class svm Error Rate=1.000000 [1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1] acc = 1.0

ocgan epoch 2

d_loss, g_loss ,xx,yy= train1(GAN, G, D, epochs=100, verbose=True)
new_data = gen(GAN, G, D, times = 10,verbose=True)
test_y_predicted = clf.predict(new_data)
print(test_y_predicted,'error = ',np.mean(test_y_predicted))

HBox(children=(IntProgress(value=0), HTML(value=’’)))

The 0 times epoch one class svm Error Rate=1.000000 The 1 times epoch one class svm Error Rate=1.000000 The 2 times epoch one class svm Error Rate=1.000000 The 3 times epoch one class svm Error Rate=1.000000 The 4 times epoch one class svm Error Rate=1.000000 The 5 times epoch one class svm Error Rate=1.000000 The 6 times epoch one class svm Error Rate=1.000000 The 7 times epoch one class svm Error Rate=1.000000 The 8 times epoch one class svm Error Rate=1.000000 The 9 times epoch one class svm Error Rate=1.000000 The 10 times epoch one class svm Error Rate=1.000000 The 11 times epoch one class svm Error Rate=1.000000 The 12 times epoch one class svm Error Rate=1.000000 The 13 times epoch one class svm Error Rate=1.000000 The 14 times epoch one class svm Error Rate=1.000000 The 15 times epoch one class svm Error Rate=1.000000 The 16 times epoch one class svm Error Rate=1.000000 The 17 times epoch one class svm Error Rate=1.000000 The 18 times epoch one class svm Error Rate=1.000000 The 19 times epoch one class svm Error Rate=1.000000 The 20 times epoch one class svm Error Rate=1.000000 The 21 times epoch one class svm Error Rate=1.000000 The 22 times epoch one class svm Error Rate=1.000000 The 23 times epoch one class svm Error Rate=1.000000 The 24 times epoch one class svm Error Rate=1.000000 The 25 times epoch one class svm Error Rate=1.000000 The 26 times epoch one class svm Error Rate=1.000000 The 27 times epoch one class svm Error Rate=1.000000 The 28 times epoch one class svm Error Rate=1.000000 The 29 times epoch one class svm Error Rate=1.000000 The 30 times epoch one class svm Error Rate=1.000000 The 31 times epoch one class svm Error Rate=1.000000 The 32 times epoch one class svm Error Rate=1.000000 The 33 times epoch one class svm Error Rate=1.000000 The 34 times epoch one class svm Error Rate=1.000000 The 35 times epoch one class svm Error Rate=1.000000 The 36 times epoch one class svm Error Rate=1.000000 The 37 times epoch one class svm Error Rate=1.000000 The 38 times epoch one class svm Error Rate=1.000000 The 39 times epoch one class svm Error Rate=1.000000 The 40 times epoch one class svm Error Rate=1.000000 The 41 times epoch one class svm Error Rate=1.000000 The 42 times epoch one class svm Error Rate=1.000000 The 43 times epoch one class svm Error Rate=1.000000 The 44 times epoch one class svm Error Rate=1.000000 The 45 times epoch one class svm Error Rate=1.000000 The 46 times epoch one class svm Error Rate=1.000000 The 47 times epoch one class svm Error Rate=1.000000 The 48 times epoch one class svm Error Rate=1.000000 The 49 times epoch one class svm Error Rate=1.000000 The 50 times epoch one class svm Error Rate=1.000000 The 51 times epoch one class svm Error Rate=1.000000 The 52 times epoch one class svm Error Rate=1.000000 The 53 times epoch one class svm Error Rate=1.000000 The 54 times epoch one class svm Error Rate=1.000000 The 55 times epoch one class svm Error Rate=1.000000 The 56 times epoch one class svm Error Rate=1.000000 The 57 times epoch one class svm Error Rate=1.000000 The 58 times epoch one class svm Error Rate=1.000000 The 59 times epoch one class svm Error Rate=1.000000 The 60 times epoch one class svm Error Rate=1.000000 The 61 times epoch one class svm Error Rate=1.000000 The 62 times epoch one class svm Error Rate=1.000000 The 63 times epoch one class svm Error Rate=1.000000 The 64 times epoch one class svm Error Rate=1.000000 The 65 times epoch one class svm Error Rate=1.000000 Epoch #66: Generative Loss: 4.731652736663818, Discriminative Loss: 0.021196838468313217 The 66 times epoch one class svm Error Rate=1.000000 The 67 times epoch one class svm Error Rate=1.000000 The 68 times epoch one class svm Error Rate=1.000000 The 69 times epoch one class svm Error Rate=1.000000 The 70 times epoch one class svm Error Rate=1.000000 The 71 times epoch one class svm Error Rate=1.000000 The 72 times epoch one class svm Error Rate=1.000000 The 73 times epoch one class svm Error Rate=1.000000 The 74 times epoch one class svm Error Rate=1.000000 The 75 times epoch one class svm Error Rate=1.000000 The 76 times epoch one class svm Error Rate=1.000000 The 77 times epoch one class svm Error Rate=1.000000 The 78 times epoch one class svm Error Rate=1.000000 The 79 times epoch one class svm Error Rate=1.000000 The 80 times epoch one class svm Error Rate=1.000000 The 81 times epoch one class svm Error Rate=1.000000 The 82 times epoch one class svm Error Rate=1.000000 The 83 times epoch one class svm Error Rate=0.978261 The 84 times epoch one class svm Error Rate=0.913043 The 85 times epoch one class svm Error Rate=0.934783 The 86 times epoch one class svm Error Rate=0.869565 The 87 times epoch one class svm Error Rate=0.782609 The 88 times epoch one class svm Error Rate=0.782609 The 89 times epoch one class svm Error Rate=0.673913 The 90 times epoch one class svm Error Rate=0.478261 The 91 times epoch one class svm Error Rate=0.478261 The 92 times epoch one class svm Error Rate=0.347826 The 93 times epoch one class svm Error Rate=0.521739 The 94 times epoch one class svm Error Rate=0.152174 The 95 times epoch one class svm Error Rate=0.217391 The 96 times epoch one class svm Error Rate=0.195652 The 97 times epoch one class svm Error Rate=0.130435 The 98 times epoch one class svm Error Rate=0.130435 The 99 times epoch one class svm Error Rate=0.217391

The 0 times generator one class svm Error Rate=0.043478 The 1 times generator one class svm Error Rate=0.086957 The 2 times generator one class svm Error Rate=0.043478 The 3 times generator one class svm Error Rate=0.108696 The 4 times generator one class svm Error Rate=0.108696 The 5 times generator one class svm Error Rate=0.021739 The 6 times generator one class svm Error Rate=0.065217 The 7 times generator one class svm Error Rate=0.021739 The 8 times generator one class svm Error Rate=0.086957 The 9 times generator one class svm Error Rate=0.043478 [1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1] error = 1.0

ocgan epoch 3

d_loss, g_loss ,xx,yy= train1(GAN, G, D, epochs=10, verbose=True)
new_data = gen(GAN, G, D, times = 10,verbose=True)
test_y_predicted = clf.predict(new_data)
print(test_y_predicted,'acc = ',np.mean(test_y_predicted))

HBox(children=(IntProgress(value=0, max=10), HTML(value=’’)))

The 0 times epoch one class svm Error Rate=0.065217 The 1 times epoch one class svm Error Rate=0.152174 The 2 times epoch one class svm Error Rate=0.000000 The 3 times epoch one class svm Error Rate=0.065217 The 4 times epoch one class svm Error Rate=0.021739 The 5 times epoch one class svm Error Rate=0.021739 The 6 times epoch one class svm Error Rate=0.021739 The 7 times epoch one class svm Error Rate=0.021739 The 8 times epoch one class svm Error Rate=0.000000 The 9 times epoch one class svm Error Rate=0.000000

The 0 times generator one class svm Error Rate=0.000000 The 1 times generator one class svm Error Rate=0.000000 The 2 times generator one class svm Error Rate=0.000000 The 3 times generator one class svm Error Rate=0.021739 The 4 times generator one class svm Error Rate=0.021739 The 5 times generator one class svm Error Rate=0.000000 The 6 times generator one class svm Error Rate=0.000000 The 7 times generator one class svm Error Rate=0.000000 The 8 times generator one class svm Error Rate=0.000000 The 9 times generator one class svm Error Rate=0.000000 [1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1] acc = 1.0