Using TensorFlow backend.

Write an autoencoder to do the fraud detection using Keras.

Write an autoencoder to do the fraud detection using Keras.

Make inputs

import numpy as np

np.random.seed(1) # reproducibility

import pandas as pd

df = pd.read_csv("data/creditcard.csv")

df.head()

	Time	V1	V2	V3	V4	V5	V6	V7	V8	V9	...	V21	V22	V23	V24	V25	V26	V27	V28	Amount	Class
0	0.0	-1.359807	-0.072781	2.536347	1.378155	-0.338321	0.462388	0.239599	0.098698	0.363787	...	-0.018307	0.277838	-0.110474	0.066928	0.128539	-0.189115	0.133558	-0.021053	149.62	0
1	0.0	1.191857	0.266151	0.166480	0.448154	0.060018	-0.082361	-0.078803	0.085102	-0.255425	...	-0.225775	-0.638672	0.101288	-0.339846	0.167170	0.125895	-0.008983	0.014724	2.69	0
2	1.0	-1.358354	-1.340163	1.773209	0.379780	-0.503198	1.800499	0.791461	0.247676	-1.514654	...	0.247998	0.771679	0.909412	-0.689281	-0.327642	-0.139097	-0.055353	-0.059752	378.66	0
3	1.0	-0.966272	-0.185226	1.792993	-0.863291	-0.010309	1.247203	0.237609	0.377436	-1.387024	...	-0.108300	0.005274	-0.190321	-1.175575	0.647376	-0.221929	0.062723	0.061458	123.50	0
4	2.0	-1.158233	0.877737	1.548718	0.403034	-0.407193	0.095921	0.592941	-0.270533	0.817739	...	-0.009431	0.798278	-0.137458	0.141267	-0.206010	0.502292	0.219422	0.215153	69.99	0

5 rows ¡Á 31 columns

df.shape

(284807, 31)

y = df['Class']
X = df.drop(['Time', 'Class'], axis = 1)

from sklearn.model_selection import train_test_split

# ?train_test_split
X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.1, random_state=42)

Define and train model

build_model[source]

build_model(input_dim=29, outer_layer_dim=14, inner_layer_dim=7)

# ?autoencoder.compile

autoencoder.compile(optimizer='adam', loss='mean_squared_error', metrics=['accuracy'])
history = autoencoder.fit(X_train, X_train, epochs = 10, batch_size = 32, validation_split=0.2)

Train on 205060 samples, validate on 51266 samples
Epoch 1/10
205060/205060 [==============================] - 10s 47us/step - loss: 58.7521 - accuracy: 0.8417 - val_loss: 0.9116 - val_accuracy: 0.8536
Epoch 2/10
205060/205060 [==============================] - 10s 50us/step - loss: 2.4403 - accuracy: 0.8613 - val_loss: 0.7706 - val_accuracy: 0.8786
Epoch 3/10
205060/205060 [==============================] - 9s 45us/step - loss: 1.6772 - accuracy: 0.8856 - val_loss: 0.7180 - val_accuracy: 0.8955
Epoch 4/10
205060/205060 [==============================] - 9s 44us/step - loss: 1.3769 - accuracy: 0.8986 - val_loss: 0.9737 - val_accuracy: 0.9111
Epoch 5/10
205060/205060 [==============================] - 9s 45us/step - loss: 1.7829 - accuracy: 0.9004 - val_loss: 0.7955 - val_accuracy: 0.9119
Epoch 6/10
205060/205060 [==============================] - 9s 45us/step - loss: 1.2445 - accuracy: 0.9101 - val_loss: 0.6326 - val_accuracy: 0.9171
Epoch 7/10
205060/205060 [==============================] - 10s 47us/step - loss: 1.5845 - accuracy: 0.9157 - val_loss: 0.6526 - val_accuracy: 0.9168
Epoch 8/10
205060/205060 [==============================] - 9s 44us/step - loss: 1.2776 - accuracy: 0.9221 - val_loss: 0.7749 - val_accuracy: 0.9067
Epoch 9/10
205060/205060 [==============================] - 9s 44us/step - loss: 1.1414 - accuracy: 0.9253 - val_loss: 0.5749 - val_accuracy: 0.9349
Epoch 10/10
205060/205060 [==============================] - 9s 44us/step - loss: 1.0061 - accuracy: 0.9317 - val_loss: 0.5736 - val_accuracy: 0.9319

autoencoder.save("model/creditcard_autoencoders_model.h5")

plot_acc[source]

plot_acc(history)

plot_acc(history.history)

plot_loss[source]

plot_loss(history)

plot_loss(history.history)

from keras.models import load_model

autoencoder2 = load_model("model/creditcard_autoencoders_model.h5")
autoencoder2.__class__

keras.engine.training.Model

preds = autoencoder.predict(X_test)

for i in range(5):
    # print(i)
    if np.sum(np.square(preds[i] - np.array(X_test)[i]))/30 < 1:
        np.sum(np.square(preds[i] - np.array(X_test)[i]))/30
    else :
        1

ifelse is too slow.

y_preds  = np.where(np.sum(np.square(preds - np.array(X_test)), axis = 1)/30 < 1, 
                    np.sum(np.square(preds - np.array(X_test)), axis = 1)/30, 1)

y_preds.__class__

numpy.ndarray

y_preds.shape

(28481,)

import pyks

data = pd.DataFrame({'y': y_test, 'yhat': y_preds})

pyks.plot(data)

0.8452397152926965

0.8452397152926965

<Figure size 432x288 with 0 Axes>