In [7]:
%matplotlib inline
In [8]:
import pandas as pd
import numpy as np
from keras.utils.np_utils import to_categorical
from keras.models import Sequential
from keras.layers import Dense, Input
from keras.models import Model
from keras.wrappers.scikit_learn import KerasRegressor, KerasClassifier
from sklearn.model_selection import cross_val_score
from sklearn.model_selection import KFold
In [6]:
from sklearn.preprocessing import StandardScaler
from sklearn.pipeline import Pipeline
from utils import get_engine
from sklearn import linear_model
import matplotlib.pyplot as plt
if __name__ == '__main__':
# load dataset
table = "data_fraud_little"
engine = get_engine()
dataframe = pd.read_sql_query("select * from {table} limit 500".format(table=table),engine)
dataset = dataframe.values
print("First one row of the dataset")
print("Shape [{}]".format(dataset.shape))
print(dataset[0:2,:])
# split into input (X) and output (Y) variables
data_dimensions = 45
#first dimension is the index, must be removed!!!!
X = dataset[:, 1:data_dimensions]
Y = dataset[:, data_dimensions]
print("Fraud {}% ".format(float(np.sum(Y==1))*100.0/Y.shape[0]))
print("Total #samples:",Y.shape[0])
Y = to_categorical(Y, nb_classes=None)
input_dimensions = X.shape[1]
print("shapes: X[{}]=====Y[{}]".format(X.shape, Y.shape))
# define base mode
def baseline_model():
return logistic_regresion()
# return linear_regression()
def keras_lin_reg():
x = Input((None,input_dimensions))
y = Dense(1,activation='linear')(x)
model = Model(x,y,"Linear Regression")
model.compile(loss='mse', optimizer='sgd')
return model
def logistic_regresion():
logistic = linear_model.LogisticRegression(solver='sag', n_jobs=-1,max_iter=500)
return logistic
def linear_regression():
lr = linear_model.LinearRegression(n_jobs=-1)
return lr
def mlp_model(hidden=None,layers=1):
# create model
model = Sequential()
model.add(Dense(input_dimensions, input_dim=input_dimensions, init='normal', activation='relu'))
if hidden is not None:
for l in range(layers):
model.add(Dense(hidden))
model.add(Dense(2, init='normal', activation='softmax'))
# Compile model
model.compile(loss='binary_crossentropy', optimizer='adam')
return model
def mlp_model_wrap(layers=1):
return mlp_model(100,layers)
# fix random seed for reproducibility
seed = 7
np.random.seed(seed)
# evaluate model with standardized dataset
estimators = []
estimators.append(('standardize', StandardScaler()))
# estimators.append(('mlp', KerasClassifier(build_fn=mlp_model, nb_epoch=100, batch_size=10000, verbose=1)))
estimators.append(('mlp', KerasClassifier(build_fn=mlp_model_wrap, nb_epoch=100, batch_size=10000, verbose=0)))
# estimators.append(('liner reg', KerasClassifier(build_fn=keras_lin_reg, nb_epoch=100, batch_size=100000, verbose=1)))
# estimators.append(('linear_reg', baseline_model()))
pipeline = Pipeline(estimators)
pipeline.set_params(mlp__layers=2)
kfold = KFold(n_splits=2, random_state=seed)
results = cross_val_score(pipeline, X, Y, cv=kfold, scoring='roc_auc',n_jobs=1)
print("Results:", results)
print("Results: %.24f (%.24f) ROC" % (results.mean(), results.std()))
print(pipeline)
plt.errorbar([0], [results.mean()], np.array(results.std()))
plt.title(
'Cross Validation')
plt.xlabel('Model')
plt.ylabel('AUC')
plt.axis('tight')
plt.show()
First one row of the dataset
Shape [(500, 47)]
[[ 4.75200320e+07 7.49524000e+05 1.38868315e+18 6.89620000e+04
5.83500000e+01 1.77484800e+06 5.77500000e+04 1.87362000e+05
3.67000000e+02 1.62381800e+06 1.68885100e+06 5.20000000e+02
5.00000000e+00 0.00000000e+00 0.00000000e+00 8.00000000e+00
1.00000000e+00 0.00000000e+00 4.00000000e+00 0.00000000e+00
1.00000000e+00 1.00000000e+02 0.00000000e+00 0.00000000e+00
0.00000000e+00 5.60000000e+01 1.00000000e+00 1.00000000e+00
0.00000000e+00 1.00000000e+00 3.00000000e+00 0.00000000e+00
1.90000000e+01 1.84000000e+03 8.26000000e+02 6.10000000e+07
5.83500000e+01 5.13265000e+03 7.50000000e+03 7.50000000e+03
2.18985000e+03 -9.22337204e+18 1.38602880e+18 1.47048000e+03
1.37419200e+18 0.00000000e+00 -9.22337204e+18]
[ 3.88424760e+07 6.72120000e+05 1.39431354e+18 1.00165000e+05
1.49900000e+01 4.20536000e+05 1.32586000e+05 0.00000000e+00
3.67000000e+02 6.71779000e+05 1.49575000e+06 6.12000000e+02
1.00000000e+00 8.00000000e+00 0.00000000e+00 9.00000000e+00
2.00000000e+00 4.80000000e+01 1.00000000e+00 4.00000000e+00
2.00000000e+00 1.00000000e+02 0.00000000e+00 5.00000000e+00
1.42000000e+02 9.80000000e+01 1.00000000e+00 1.00000000e+00
0.00000000e+00 3.00000000e+00 3.00000000e+00 0.00000000e+00
1.90000000e+01 8.28300000e+03 8.26000000e+02 6.10000000e+07
0.00000000e+00 -3.40400000e+01 1.00000000e+02 1.00000000e+02
1.34040000e+02 1.34904960e+18 1.39173120e+18 5.72400000e+01
1.38101760e+18 0.00000000e+00 -9.22337204e+18]]
Fraud 15.4%
('Total #samples:', 500)
shapes: X[(500, 44)]=====Y[(500, 2)]
('Results:', array([ 0.81347979, 0.90068653]))
Results: 0.857083160019660894590743 (0.043603374541113071316545) ROC
Pipeline(steps=[('standardize', StandardScaler(copy=True, with_mean=True, with_std=True)), ('mlp', <keras.wrappers.scikit_learn.KerasClassifier object at 0x7f205195ce50>)])
In [10]:
pipeline.fit(X,Y)
pipeline.predict_proba(X)
Out[10]:
array([[ 1.00000000e+00, 2.34180564e-10],
[ 9.91887629e-01, 8.11230578e-03],
[ 3.39784436e-02, 9.66021597e-01],
[ 9.93469596e-01, 6.53044274e-03],
[ 2.36431565e-02, 9.76356864e-01],
[ 1.00000000e+00, 3.87060306e-12],
[ 1.00000000e+00, 1.53236132e-10],
[ 1.00000000e+00, 2.31401631e-09],
[ 9.99999285e-01, 6.85733085e-07],
[ 9.99383569e-01, 6.16431120e-04],
[ 1.00000000e+00, 4.76656359e-09],
[ 7.88428187e-01, 2.11571753e-01],
[ 1.00000000e+00, 1.16372743e-12],
[ 1.00000000e+00, 3.79162035e-09],
[ 1.00000000e+00, 9.22906640e-09],
[ 5.39086908e-02, 9.46091354e-01],
[ 1.00000000e+00, 2.20846865e-11],
[ 9.99942660e-01, 5.73005382e-05],
[ 8.28953460e-04, 9.99171019e-01],
[ 1.00000000e+00, 3.23459287e-10],
[ 9.84769940e-01, 1.52300727e-02],
[ 9.95903194e-01, 4.09686007e-03],
[ 2.41316020e-01, 7.58683980e-01],
[ 1.00000000e+00, 1.23799374e-10],
[ 9.99999046e-01, 9.48769241e-07],
[ 9.98486400e-01, 1.51351886e-03],
[ 3.29264365e-02, 9.67073560e-01],
[ 1.00000000e+00, 2.77303770e-11],
[ 1.00000000e+00, 9.47222745e-10],
[ 1.00000000e+00, 2.31689556e-09],
[ 1.04176201e-01, 8.95823836e-01],
[ 1.00000000e+00, 1.52887477e-11],
[ 9.99998093e-01, 1.91769664e-06],
[ 1.00000000e+00, 1.71420229e-08],
[ 1.00000000e+00, 6.95677731e-15],
[ 1.00000000e+00, 1.45227580e-13],
[ 9.90409136e-01, 9.59082227e-03],
[ 9.99999642e-01, 3.36186446e-07],
[ 1.00000000e+00, 4.86535923e-10],
[ 9.99994040e-01, 5.98493898e-06],
[ 9.80577946e-01, 1.94220077e-02],
[ 9.99812186e-01, 1.87792379e-04],
[ 9.99997854e-01, 2.19989715e-06],
[ 9.81814027e-01, 1.81859843e-02],
[ 1.00000000e+00, 7.63322077e-13],
[ 9.99983788e-01, 1.62487377e-05],
[ 9.23395064e-03, 9.90766048e-01],
[ 1.00000000e+00, 1.66607460e-17],
[ 1.07582845e-02, 9.89241779e-01],
[ 9.69576478e-01, 3.04235090e-02],
[ 1.00000000e+00, 2.19642704e-13],
[ 6.36708736e-03, 9.93632913e-01],
[ 1.00000000e+00, 6.46676224e-09],
[ 9.99999881e-01, 7.01714384e-08],
[ 1.00000000e+00, 4.50736670e-10],
[ 1.00000000e+00, 1.27155646e-08],
[ 1.63780116e-02, 9.83622015e-01],
[ 1.00000000e+00, 1.00191123e-11],
[ 1.00000000e+00, 8.86633100e-10],
[ 1.00000000e+00, 8.52195647e-09],
[ 1.00000000e+00, 2.05669912e-10],
[ 1.00000000e+00, 1.21613803e-08],
[ 1.25748190e-14, 1.00000000e+00],
[ 1.00000000e+00, 3.64715333e-08],
[ 2.40057912e-02, 9.75994229e-01],
[ 1.00000000e+00, 1.60403895e-11],
[ 1.00000000e+00, 6.63837518e-09],
[ 1.00000000e+00, 6.92635021e-23],
[ 1.00000000e+00, 9.06929566e-12],
[ 9.97506440e-01, 2.49362085e-03],
[ 1.00000000e+00, 1.60579267e-15],
[ 1.00000000e+00, 1.42051582e-09],
[ 9.99998808e-01, 1.17872582e-06],
[ 9.99998450e-01, 1.50924689e-06],
[ 9.99999285e-01, 7.70872020e-07],
[ 9.99986649e-01, 1.33832445e-05],
[ 1.00000000e+00, 4.13898995e-08],
[ 9.95781660e-01, 4.21831524e-03],
[ 1.00000000e+00, 6.33169683e-10],
[ 9.86253858e-01, 1.37461079e-02],
[ 1.00000000e+00, 2.20828262e-08],
[ 8.97186756e-01, 1.02813236e-01],
[ 1.76940709e-01, 8.23059320e-01],
[ 1.09679217e-03, 9.98903155e-01],
[ 9.99996662e-01, 3.37050460e-06],
[ 1.00000000e+00, 9.31325683e-09],
[ 1.00000000e+00, 3.99776342e-14],
[ 1.00000000e+00, 9.37404154e-10],
[ 9.99706209e-01, 2.93755991e-04],
[ 3.04562156e-04, 9.99695420e-01],
[ 4.57073515e-03, 9.95429277e-01],
[ 1.00000000e+00, 1.28642404e-13],
[ 9.97191012e-01, 2.80898274e-03],
[ 7.08220541e-01, 2.91779488e-01],
[ 9.99931097e-01, 6.88531582e-05],
[ 9.99686241e-01, 3.13809665e-04],
[ 9.99980688e-01, 1.93021224e-05],
[ 9.99998093e-01, 1.89938442e-06],
[ 1.00000000e+00, 5.15498998e-11],
[ 1.00000000e+00, 6.63272855e-17],
[ 1.00000000e+00, 2.53280046e-14],
[ 9.99999881e-01, 5.99793850e-08],
[ 4.03057486e-02, 9.59694266e-01],
[ 7.21912384e-02, 9.27808702e-01],
[ 9.90763843e-01, 9.23623797e-03],
[ 1.00000000e+00, 4.53907392e-13],
[ 1.00000000e+00, 1.76316351e-12],
[ 1.20003580e-03, 9.98799920e-01],
[ 9.99999046e-01, 1.00636862e-06],
[ 1.00000000e+00, 8.66220182e-13],
[ 1.00000000e+00, 4.14777590e-09],
[ 5.70077375e-02, 9.42992210e-01],
[ 6.56816829e-03, 9.93431866e-01],
[ 9.80510414e-01, 1.94895752e-02],
[ 8.38613790e-03, 9.91613865e-01],
[ 9.94996548e-01, 5.00346394e-03],
[ 1.00000000e+00, 1.21470771e-08],
[ 9.99998331e-01, 1.63975039e-06],
[ 1.00000000e+00, 2.70132823e-08],
[ 1.00000000e+00, 4.06455121e-12],
[ 9.55926112e-07, 9.99999046e-01],
[ 1.00000000e+00, 2.13286458e-10],
[ 8.59646261e-01, 1.40353769e-01],
[ 1.00000000e+00, 2.81609158e-09],
[ 9.99999881e-01, 1.05152310e-07],
[ 1.00000000e+00, 1.46881192e-13],
[ 3.37449014e-02, 9.66255069e-01],
[ 9.99992847e-01, 7.12007704e-06],
[ 9.99783099e-01, 2.16947446e-04],
[ 1.00000000e+00, 7.23513638e-10],
[ 1.00000000e+00, 8.29704252e-13],
[ 9.99408722e-01, 5.91335178e-04],
[ 1.00000000e+00, 2.38379316e-10],
[ 9.99991775e-01, 8.21493904e-06],
[ 9.99962568e-01, 3.74736010e-05],
[ 1.00000000e+00, 4.40679741e-12],
[ 9.99985576e-01, 1.44810192e-05],
[ 1.72896519e-01, 8.27103496e-01],
[ 1.07187415e-02, 9.89281178e-01],
[ 3.44011234e-03, 9.96559918e-01],
[ 9.99999404e-01, 5.57424244e-07],
[ 9.99082327e-01, 9.17693775e-04],
[ 9.99998093e-01, 1.90471553e-06],
[ 1.00000000e+00, 1.61352126e-10],
[ 1.00000000e+00, 9.42395359e-13],
[ 9.92796123e-01, 7.20382063e-03],
[ 1.00000000e+00, 1.57830584e-08],
[ 1.00000000e+00, 1.56971987e-15],
[ 9.98684943e-01, 1.31505390e-03],
[ 9.99996662e-01, 3.33628896e-06],
[ 9.44218397e-01, 5.57815805e-02],
[ 6.41771108e-02, 9.35822904e-01],
[ 1.00000000e+00, 1.60415059e-09],
[ 9.99808729e-01, 1.91253770e-04],
[ 4.25532181e-03, 9.95744646e-01],
[ 1.00000000e+00, 1.39949146e-08],
[ 1.00000000e+00, 4.36797489e-13],
[ 1.00000000e+00, 3.79696968e-12],
[ 9.57793593e-01, 4.22064811e-02],
[ 9.99999642e-01, 3.16412724e-07],
[ 1.00000000e+00, 1.00767016e-11],
[ 9.99960661e-01, 3.93816263e-05],
[ 1.00000000e+00, 2.37325784e-11],
[ 1.00000000e+00, 3.95568662e-08],
[ 9.99851346e-01, 1.48610998e-04],
[ 4.07182937e-03, 9.95928109e-01],
[ 1.00000000e+00, 7.56628926e-11],
[ 1.00000000e+00, 2.60296038e-12],
[ 9.99999762e-01, 2.92505348e-07],
[ 5.38654141e-02, 9.46134567e-01],
[ 9.99996662e-01, 3.28740680e-06],
[ 9.99987245e-01, 1.27193516e-05],
[ 9.98968124e-01, 1.03189040e-03],
[ 9.99384999e-01, 6.14952936e-04],
[ 9.99984622e-01, 1.54268946e-05],
[ 1.00000000e+00, 4.19389801e-10],
[ 1.00000000e+00, 5.29668377e-13],
[ 1.00000000e+00, 4.92776718e-13],
[ 9.99994516e-01, 5.52847678e-06],
[ 8.55198920e-01, 1.44801095e-01],
[ 1.00000000e+00, 2.10576931e-14],
[ 1.00000000e+00, 3.18217772e-08],
[ 1.43963620e-02, 9.85603690e-01],
[ 9.99999762e-01, 1.94984779e-07],
[ 1.00000000e+00, 3.51057643e-13],
[ 1.00000000e+00, 4.00819135e-11],
[ 2.88269226e-03, 9.97117281e-01],
[ 9.99833465e-01, 1.66493468e-04],
[ 9.97124732e-01, 2.87530874e-03],
[ 8.72349977e-01, 1.27650082e-01],
[ 9.99991417e-01, 8.60157888e-06],
[ 7.37464754e-03, 9.92625296e-01],
[ 1.00000000e+00, 1.25247078e-12],
[ 4.84207422e-02, 9.51579273e-01],
[ 1.00000000e+00, 3.04910402e-10],
[ 1.00000000e+00, 6.94036428e-09],
[ 2.64786091e-03, 9.97352123e-01],
[ 1.00000000e+00, 4.53861526e-09],
[ 5.93228033e-03, 9.94067669e-01],
[ 5.09469677e-03, 9.94905353e-01],
[ 1.58681087e-02, 9.84131873e-01],
[ 1.00000000e+00, 2.35901472e-11],
[ 9.98698950e-01, 1.30096904e-03],
[ 1.00000000e+00, 1.73976332e-15],
[ 9.99999046e-01, 9.50754668e-07],
[ 9.99940276e-01, 5.97653270e-05],
[ 9.93408799e-01, 6.59123575e-03],
[ 9.81047809e-01, 1.89521648e-02],
[ 9.97092128e-01, 2.90781143e-03],
[ 9.99680638e-01, 3.19412415e-04],
[ 9.99859691e-01, 1.40344899e-04],
[ 1.00000000e+00, 5.84735524e-11],
[ 1.61855214e-03, 9.98381495e-01],
[ 9.62868571e-01, 3.71314920e-02],
[ 1.00000000e+00, 2.08608104e-14],
[ 1.00000000e+00, 1.09285958e-09],
[ 9.99992728e-01, 7.29761132e-06],
[ 9.99833822e-01, 1.66190832e-04],
[ 9.99889731e-01, 1.10314133e-04],
[ 1.00000000e+00, 1.71272085e-09],
[ 9.99999881e-01, 7.06925292e-08],
[ 9.99997973e-01, 2.05446054e-06],
[ 9.98331606e-01, 1.66838092e-03],
[ 9.01462361e-02, 9.09853756e-01],
[ 1.00000000e+00, 2.03968873e-11],
[ 9.99999642e-01, 4.16047214e-07],
[ 1.00000000e+00, 6.58931620e-09],
[ 1.34354401e-02, 9.86564636e-01],
[ 1.00000000e+00, 3.60211752e-13],
[ 9.99999762e-01, 2.49756340e-07],
[ 1.00000000e+00, 1.35197721e-13],
[ 9.99999881e-01, 1.69840717e-07],
[ 9.92859364e-01, 7.14065880e-03],
[ 2.08264566e-03, 9.97917354e-01],
[ 1.00000000e+00, 1.41497845e-08],
[ 1.00000000e+00, 5.73373987e-12],
[ 9.99998808e-01, 1.16822184e-06],
[ 9.99997735e-01, 2.27003238e-06],
[ 9.29973722e-02, 9.07002628e-01],
[ 1.00000000e+00, 1.39875730e-12],
[ 1.13576057e-03, 9.98864293e-01],
[ 1.00000000e+00, 1.89494962e-08],
[ 1.00000000e+00, 1.92475302e-09],
[ 4.70994785e-03, 9.95290041e-01],
[ 9.92927372e-01, 7.07262661e-03],
[ 1.00000000e+00, 1.29170181e-11],
[ 9.99972701e-01, 2.73071892e-05],
[ 1.00000000e+00, 1.58940319e-10],
[ 2.27149867e-05, 9.99977231e-01],
[ 9.96167839e-01, 3.83209181e-03],
[ 1.00000000e+00, 4.36041692e-09],
[ 1.00000000e+00, 2.74426033e-12],
[ 9.99861360e-01, 1.38627671e-04],
[ 9.93383944e-01, 6.61600986e-03],
[ 9.99962687e-01, 3.73507683e-05],
[ 9.89573896e-01, 1.04260892e-02],
[ 9.95230973e-01, 4.76895319e-03],
[ 3.28875799e-03, 9.96711254e-01],
[ 9.99989390e-01, 1.06620155e-05],
[ 1.00000000e+00, 3.27083666e-10],
[ 1.00000000e+00, 1.40590303e-10],
[ 1.00000000e+00, 4.06158343e-08],
[ 9.99950886e-01, 4.91197716e-05],
[ 1.00000000e+00, 2.78245114e-12],
[ 3.64263617e-02, 9.63573635e-01],
[ 9.99504328e-01, 4.95659944e-04],
[ 9.99958992e-01, 4.09542845e-05],
[ 5.09787817e-03, 9.94902134e-01],
[ 9.99996662e-01, 3.33001708e-06],
[ 9.99946713e-01, 5.33106031e-05],
[ 9.99999881e-01, 1.18181617e-07],
[ 1.00000000e+00, 1.52418060e-08],
[ 9.95813191e-01, 4.18681838e-03],
[ 1.00000000e+00, 3.47629503e-09],
[ 9.90888953e-01, 9.11108777e-03],
[ 6.02732822e-02, 9.39726770e-01],
[ 9.88411605e-01, 1.15883788e-02],
[ 1.00000000e+00, 5.73068748e-09],
[ 1.00000000e+00, 1.18141616e-11],
[ 9.99889016e-01, 1.10962108e-04],
[ 8.46215979e-12, 1.00000000e+00],
[ 9.80402231e-01, 1.95977055e-02],
[ 1.00000000e+00, 3.36166289e-13],
[ 9.99999404e-01, 5.76993216e-07],
[ 9.99999166e-01, 8.73952274e-07],
[ 7.01957121e-02, 9.29804325e-01],
[ 1.55035325e-03, 9.98449683e-01],
[ 1.00000000e+00, 1.01377996e-10],
[ 1.00000000e+00, 1.82100793e-11],
[ 1.00000000e+00, 3.25540463e-12],
[ 1.00000000e+00, 1.29711741e-09],
[ 1.00573397e-03, 9.98994291e-01],
[ 1.00000000e+00, 2.57344857e-09],
[ 1.00000000e+00, 1.24967476e-11],
[ 1.00000000e+00, 2.56309640e-10],
[ 1.00000000e+00, 1.98705630e-09],
[ 1.00000000e+00, 1.15298537e-12],
[ 9.78253484e-01, 2.17464920e-02],
[ 4.28757891e-02, 9.57124233e-01],
[ 9.99998808e-01, 1.17485843e-06],
[ 9.99402642e-01, 5.97391801e-04],
[ 1.95776783e-02, 9.80422318e-01],
[ 1.00000000e+00, 1.96990602e-08],
[ 9.99492049e-01, 5.07909805e-04],
[ 9.99714792e-01, 2.85236340e-04],
[ 9.99996662e-01, 3.28271381e-06],
[ 1.00000000e+00, 5.51129758e-12],
[ 9.99999046e-01, 9.24383016e-07],
[ 9.87278342e-01, 1.27216848e-02],
[ 1.00000000e+00, 1.86039725e-12],
[ 1.00000000e+00, 3.71836485e-11],
[ 9.99976993e-01, 2.30459718e-05],
[ 1.00000000e+00, 1.41928938e-12],
[ 1.00000000e+00, 8.31741134e-11],
[ 1.00000000e+00, 1.97591249e-11],
[ 9.97482479e-01, 2.51754629e-03],
[ 1.06231123e-02, 9.89376903e-01],
[ 9.99999642e-01, 3.30791011e-07],
[ 1.00000000e+00, 7.76545939e-09],
[ 1.00000000e+00, 1.67049721e-10],
[ 9.84578311e-01, 1.54217137e-02],
[ 1.00000000e+00, 1.47617734e-11],
[ 1.00000000e+00, 5.70436719e-11],
[ 9.99997497e-01, 2.45918272e-06],
[ 1.00000000e+00, 1.26389411e-19],
[ 1.00000000e+00, 5.87360091e-11],
[ 1.00000000e+00, 7.72278617e-13],
[ 9.99630213e-01, 3.69710295e-04],
[ 1.00000000e+00, 1.77255849e-11],
[ 9.99999166e-01, 8.40471671e-07],
[ 9.99994040e-01, 5.93373306e-06],
[ 1.18872046e-01, 8.81127954e-01],
[ 1.00000000e+00, 1.08109224e-16],
[ 1.00000000e+00, 1.63808602e-11],
[ 9.99912739e-01, 8.72182572e-05],
[ 1.43260583e-02, 9.85673964e-01],
[ 9.97038722e-01, 2.96131335e-03],
[ 9.99999881e-01, 1.69614111e-07],
[ 1.00000000e+00, 2.49660764e-10],
[ 9.99993920e-01, 6.13029169e-06],
[ 1.00000000e+00, 1.34484035e-09],
[ 9.88722205e-01, 1.12777539e-02],
[ 9.99999881e-01, 1.55356730e-07],
[ 9.99977946e-01, 2.20534985e-05],
[ 9.99928713e-01, 7.13230475e-05],
[ 9.87058580e-01, 1.29414713e-02],
[ 9.99997020e-01, 2.96485427e-06],
[ 1.00000000e+00, 3.22561995e-08],
[ 1.00000000e+00, 1.61426499e-08],
[ 9.83405352e-01, 1.65946726e-02],
[ 1.00000000e+00, 1.05576681e-16],
[ 9.99999523e-01, 4.69146840e-07],
[ 9.99973178e-01, 2.68013137e-05],
[ 9.99986172e-01, 1.38578143e-05],
[ 1.00000000e+00, 3.21390809e-12],
[ 1.00000000e+00, 1.56338359e-10],
[ 9.99761283e-01, 2.38703913e-04],
[ 9.99999881e-01, 1.56546250e-07],
[ 1.00000000e+00, 1.86048771e-14],
[ 1.00000000e+00, 2.62180749e-10],
[ 1.00000000e+00, 6.37008350e-14],
[ 9.99999762e-01, 2.67419495e-07],
[ 1.00000000e+00, 1.11990470e-14],
[ 1.00000000e+00, 3.37515849e-11],
[ 9.99952793e-01, 4.72211759e-05],
[ 9.99985337e-01, 1.46248267e-05],
[ 1.25594452e-01, 8.74405563e-01],
[ 4.21584957e-07, 9.99999523e-01],
[ 1.00000000e+00, 3.22056791e-11],
[ 1.00000000e+00, 1.35755478e-11],
[ 9.99538898e-01, 4.61149146e-04],
[ 9.99999642e-01, 3.55551975e-07],
[ 8.89149010e-01, 1.10850938e-01],
[ 1.00000000e+00, 6.93857623e-13],
[ 9.99965310e-01, 3.46793313e-05],
[ 9.96848762e-01, 3.15128500e-03],
[ 9.99999166e-01, 7.97617759e-07],
[ 1.00000000e+00, 1.45754635e-08],
[ 9.93163228e-01, 6.83681900e-03],
[ 1.00000000e+00, 3.11485594e-11],
[ 9.99845386e-01, 1.54547772e-04],
[ 1.00000000e+00, 3.62247743e-09],
[ 9.97365654e-01, 2.63437885e-03],
[ 1.44120320e-04, 9.99855876e-01],
[ 3.85690071e-02, 9.61430967e-01],
[ 1.00000000e+00, 5.76927499e-11],
[ 9.99999166e-01, 8.25952100e-07],
[ 1.00000000e+00, 3.90684152e-09],
[ 9.99843836e-01, 1.56178678e-04],
[ 9.99999881e-01, 8.82278215e-08],
[ 1.00000000e+00, 4.68134083e-08],
[ 1.00000000e+00, 2.11254360e-18],
[ 9.94185507e-01, 5.81454067e-03],
[ 8.17705452e-01, 1.82294533e-01],
[ 1.00000000e+00, 2.43584124e-13],
[ 1.00000000e+00, 1.32014568e-08],
[ 9.39866364e-01, 6.01336919e-02],
[ 9.99996901e-01, 3.12274778e-06],
[ 1.00000000e+00, 7.58424851e-11],
[ 9.99992967e-01, 7.08533571e-06],
[ 9.87986475e-02, 9.01201308e-01],
[ 1.00000000e+00, 1.35674002e-14],
[ 1.00000000e+00, 2.75164957e-09],
[ 9.99999762e-01, 1.86582795e-07],
[ 9.99910355e-01, 8.96781421e-05],
[ 9.50948954e-01, 4.90510426e-02],
[ 9.99568403e-01, 4.31627588e-04],
[ 2.22208910e-03, 9.97777879e-01],
[ 1.00000000e+00, 2.20105290e-09],
[ 1.00000000e+00, 2.76623258e-08],
[ 1.00000000e+00, 9.05081010e-11],
[ 9.99996662e-01, 3.29289810e-06],
[ 1.00000000e+00, 1.11779275e-09],
[ 9.99998450e-01, 1.51502661e-06],
[ 1.00000000e+00, 3.91953525e-10],
[ 9.99835134e-01, 1.64795580e-04],
[ 1.00000000e+00, 3.76431872e-12],
[ 1.00000000e+00, 1.47042974e-16],
[ 9.96022165e-01, 3.97783332e-03],
[ 9.99999404e-01, 6.15018564e-07],
[ 1.00000000e+00, 1.52876767e-10],
[ 3.97410691e-02, 9.60258961e-01],
[ 9.99999881e-01, 6.21851726e-08],
[ 9.99990106e-01, 9.93422236e-06],
[ 9.84703600e-01, 1.52964462e-02],
[ 9.89549756e-01, 1.04502328e-02],
[ 9.99953628e-01, 4.63392862e-05],
[ 1.00000000e+00, 1.05057438e-08],
[ 1.00000000e+00, 9.38605860e-11],
[ 9.99999046e-01, 9.12445159e-07],
[ 9.61719036e-01, 3.82809825e-02],
[ 1.00000000e+00, 2.14634510e-09],
[ 1.00000000e+00, 7.77629317e-13],
[ 1.00000000e+00, 3.96895919e-11],
[ 1.00000000e+00, 3.08183634e-09],
[ 9.99999762e-01, 1.78894311e-07],
[ 1.00000000e+00, 3.68577608e-12],
[ 1.00000000e+00, 7.71172102e-12],
[ 4.39661071e-02, 9.56033826e-01],
[ 9.99619126e-01, 3.80839774e-04],
[ 1.53205520e-03, 9.98467863e-01],
[ 9.99999642e-01, 3.00585185e-07],
[ 4.05371049e-03, 9.95946348e-01],
[ 9.99148369e-01, 8.51589022e-04],
[ 1.00000000e+00, 7.19556928e-14],
[ 4.43990871e-07, 9.99999523e-01],
[ 1.44476760e-02, 9.85552251e-01],
[ 9.94328797e-01, 5.67125482e-03],
[ 1.00000000e+00, 4.84677507e-08],
[ 1.00000000e+00, 1.20118693e-10],
[ 9.99999285e-01, 6.84032216e-07],
[ 9.99999881e-01, 7.54580611e-08],
[ 9.99988914e-01, 1.11234122e-05],
[ 9.99999642e-01, 3.20076254e-07],
[ 1.00000000e+00, 2.16526068e-08],
[ 9.99262154e-01, 7.37862254e-04],
[ 1.00000000e+00, 1.84656582e-10],
[ 9.96872666e-04, 9.99003112e-01],
[ 9.99912262e-01, 8.77213824e-05],
[ 7.48493910e-01, 2.51506120e-01],
[ 1.00000000e+00, 4.07122171e-13],
[ 1.49484214e-04, 9.99850512e-01],
[ 1.00000000e+00, 1.04806918e-11],
[ 1.00000000e+00, 2.82053181e-09],
[ 7.57128419e-03, 9.92428720e-01],
[ 9.99731719e-01, 2.68245640e-04],
[ 9.86689270e-01, 1.33107612e-02],
[ 1.00000000e+00, 4.72146378e-10],
[ 1.00000000e+00, 1.91357774e-09],
[ 1.00000000e+00, 6.13582915e-13],
[ 1.00000000e+00, 1.98619496e-10],
[ 9.99471247e-01, 5.28819335e-04],
[ 9.99993563e-01, 6.47816205e-06],
[ 9.99918461e-01, 8.15274034e-05],
[ 1.00000000e+00, 5.58172920e-13],
[ 1.00000000e+00, 7.99934874e-09],
[ 1.00000000e+00, 9.12826909e-15],
[ 1.00000000e+00, 1.24082106e-10],
[ 9.97248948e-01, 2.75107892e-03],
[ 9.99999881e-01, 9.16953553e-08],
[ 1.00000000e+00, 3.52129490e-08],
[ 9.44486618e-01, 5.55133894e-02],
[ 9.99999642e-01, 3.15677283e-07],
[ 1.00000000e+00, 1.24398214e-10],
[ 9.96027708e-01, 3.97226214e-03],
[ 9.97628033e-01, 2.37197289e-03],
[ 1.00000000e+00, 2.77241229e-12],
[ 9.99999762e-01, 2.21414126e-07],
[ 9.90820706e-01, 9.17936303e-03],
[ 9.99999881e-01, 6.66470257e-08],
[ 9.99906540e-01, 9.34865602e-05],
[ 1.00000000e+00, 5.36628955e-13],
[ 1.00000000e+00, 1.01701821e-10],
[ 1.00000000e+00, 7.31026406e-10],
[ 1.02546960e-01, 8.97453070e-01],
[ 1.00000000e+00, 2.66374094e-17],
[ 9.79468942e-01, 2.05310099e-02],
[ 1.00000000e+00, 5.63526927e-13],
[ 1.00000000e+00, 3.22298075e-16],
[ 1.00000000e+00, 1.84641347e-11]], dtype=float32)
In [37]:
import model_export.py
NameErrorTraceback (most recent call last)
<ipython-input-37-1ce8a58815b3> in <module>()
----> 1 import model_export.py
/home/botty/Documents/CCFD/ccfd_dnn/model_export.py in <module>()
1 import h5py
2 import numpy as np
----> 3 from model import *
4 import keras
5 import argparse
/home/botty/Documents/CCFD/ccfd_dnn/model.py in <module>()
270 return train[:,0:discard_id],test[:,0:discard_id],train[:,discard_id],test[:,discard_id]
271
--> 272 def chunck_seq(seq_list,seq_len=seq_len_param):
273 split_seq = map(lambda x: np.array_split(x,math.ceil(len(x)/seq_len)) if len(x)>seq_len else [x],seq_list)
274 flattened = [sequence for user_seq in split_seq for sequence in user_seq]
NameError: name 'seq_len_param' is not defined
In [25]:
0.991576
In [30]:
precision recall f1-score support
class 0 0.33 1.00 0.50 1
class 1 0.00 0.00 0.00 2
class 2 1.00 1.00 1.00 1
avg / total 0.33 0.50 0.38 4
In [32]:
Out[32]:
0.44333333333333336
In [33]:
Out[33]:
0.3325
In [ ]:
Content source: bottydim/detect-credit-card-fraud
Similar notebooks: