Tuesday, May 2, 2017

Building an Artificial Neural Network - kfold cross validation



In [1]:
import pandas as pd
import numpy as np
import matplotlib.pyplot as plt
import seaborn as sns
%matplotlib inline
import warnings
warnings.filterwarnings('ignore')
%config InlineBackend.figure_format = 'retina'

1 Import data

In [2]:
dataset = pd.read_csv("datasets/Churn_Modelling.csv")
In [3]:
dataset.head()
Out[3]:
RowNumber CustomerId Surname CreditScore Geography Gender Age Tenure Balance NumOfProducts HasCrCard IsActiveMember EstimatedSalary Exited
0 1 15634602 Hargrave 619 France Female 42 2 0.00 1 1 1 101348.88 1
1 2 15647311 Hill 608 Spain Female 41 1 83807.86 1 0 1 112542.58 0
2 3 15619304 Onio 502 France Female 42 8 159660.80 3 1 0 113931.57 1
3 4 15701354 Boni 699 France Female 39 1 0.00 2 0 0 93826.63 0
4 5 15737888 Mitchell 850 Spain Female 43 2 125510.82 1 1 1 79084.10 0
In [4]:
X = dataset.iloc[:,3:13]
y = dataset.iloc[:,13]
In [5]:
X.head()
Out[5]:
CreditScore Geography Gender Age Tenure Balance NumOfProducts HasCrCard IsActiveMember EstimatedSalary
0 619 France Female 42 2 0.00 1 1 1 101348.88
1 608 Spain Female 41 1 83807.86 1 0 1 112542.58
2 502 France Female 42 8 159660.80 3 1 0 113931.57
3 699 France Female 39 1 0.00 2 0 0 93826.63
4 850 Spain Female 43 2 125510.82 1 1 1 79084.10
In [6]:
y.head()
Out[6]:
0    1
1    0
2    1
3    0
4    0
Name: Exited, dtype: int64

2 Encode categorical independent variables

In [7]:
from sklearn.preprocessing import OneHotEncoder,LabelEncoder
label_encoder_geo = LabelEncoder()
X.iloc[:,1] = label_encoder_geo.fit_transform(X.iloc[:,1])
In [8]:
X.head()
Out[8]:
CreditScore Geography Gender Age Tenure Balance NumOfProducts HasCrCard IsActiveMember EstimatedSalary
0 619 0 Female 42 2 0.00 1 1 1 101348.88
1 608 2 Female 41 1 83807.86 1 0 1 112542.58
2 502 0 Female 42 8 159660.80 3 1 0 113931.57
3 699 0 Female 39 1 0.00 2 0 0 93826.63
4 850 2 Female 43 2 125510.82 1 1 1 79084.10
In [9]:
label_encoder_sex = LabelEncoder()
X.iloc[:,2] = label_encoder_sex.fit_transform(X.iloc[:,2])
In [10]:
X.head()
Out[10]:
CreditScore Geography Gender Age Tenure Balance NumOfProducts HasCrCard IsActiveMember EstimatedSalary
0 619 0 0 42 2 0.00 1 1 1 101348.88
1 608 2 0 41 1 83807.86 1 0 1 112542.58
2 502 0 0 42 8 159660.80 3 1 0 113931.57
3 699 0 0 39 1 0.00 2 0 0 93826.63
4 850 2 0 43 2 125510.82 1 1 1 79084.10
In [11]:
onehotencoder = OneHotEncoder(categorical_features=[1])
X = onehotencoder.fit_transform(X).toarray()
In [12]:
X
Out[12]:
array([[  1.00000000e+00,   0.00000000e+00,   0.00000000e+00, ...,
          1.00000000e+00,   1.00000000e+00,   1.01348880e+05],
       [  0.00000000e+00,   0.00000000e+00,   1.00000000e+00, ...,
          0.00000000e+00,   1.00000000e+00,   1.12542580e+05],
       [  1.00000000e+00,   0.00000000e+00,   0.00000000e+00, ...,
          1.00000000e+00,   0.00000000e+00,   1.13931570e+05],
       ..., 
       [  1.00000000e+00,   0.00000000e+00,   0.00000000e+00, ...,
          0.00000000e+00,   1.00000000e+00,   4.20855800e+04],
       [  0.00000000e+00,   1.00000000e+00,   0.00000000e+00, ...,
          1.00000000e+00,   0.00000000e+00,   9.28885200e+04],
       [  1.00000000e+00,   0.00000000e+00,   0.00000000e+00, ...,
          1.00000000e+00,   0.00000000e+00,   3.81907800e+04]])
In [13]:
X[0]
Out[13]:
array([  1.00000000e+00,   0.00000000e+00,   0.00000000e+00,
         6.19000000e+02,   0.00000000e+00,   4.20000000e+01,
         2.00000000e+00,   0.00000000e+00,   1.00000000e+00,
         1.00000000e+00,   1.00000000e+00,   1.01348880e+05])

Let's delete one categorical variable to avoid dummy variable trap

In [14]:
X = X[:,1:]
In [15]:
from sklearn.model_selection import train_test_split
In [16]:
X_train,X_test, y_train, y_test = train_test_split(X,y,test_size = 0.2, random_state = 10)

Feature scaling is compulsory for ANN

In [17]:
from sklearn.preprocessing import StandardScaler
sc = StandardScaler()
X_train = sc.fit_transform(X_train)
In [18]:
X_test = sc.fit_transform(X_test)

Let's make ANN

In [19]:
import keras

Using TensorFlow backend.
In [20]:
from keras.models import Sequential
from keras.layers import Dense

Let's do kfold cross validation

In [21]:
from keras.wrappers.scikit_learn import KerasClassifier
from sklearn.model_selection import cross_val_score
In [22]:
def build_classifier():
    classifier = Sequential()
    classifier.add(Dense(output_dim = 6, init = 'uniform',activation = 'relu', input_dim = 11))
    classifier.add(Dense(output_dim = 6, init = 'uniform',activation = 'relu'))
    classifier.add(Dense(output_dim = 1, init = 'uniform',activation = 'sigmoid'))
    classifier.compile(optimizer='adam',loss ='binary_crossentropy', metrics = ['accuracy'])
    return classifier
In [23]:
classifier = KerasClassifier(build_fn=build_classifier, batch_size = 10, epochs = 20)
In [24]:
accuracies = cross_val_score(estimator=classifier,X= X_train, y= y_train, cv = 10, n_jobs=-1)

Epoch 1/20
Epoch 1/20
Epoch 1/20
Epoch 1/20
Epoch 1/20
Epoch 1/20
Epoch 1/20
Epoch 1/20
5930/7200 [=======================>......] - ETA: 0s - loss: 0.5049 - acc: 0.7993
Epoch 2/20
7200/7200 [==============================] - 3s - loss: 0.4897 - acc: 0.7976     
Epoch 2/20

6720/7200 [===========================>..] - ETA: 0s - loss: 0.4966 - acc: 0.7976Epoch 2/20
6970/7200 [============================>.] - ETA: 0s - loss: 0.4914 - acc: 0.7989
  70/7200 [..............................] - ETA: 6s - loss: 0.5474 - acc: 0.7429Epoch 2/20
 180/7200 [..............................] - ETA: 4s - loss: 0.4203 - acc: 0.8056
 390/7200 [>.............................] - ETA: 3s - loss: 0.4445 - acc: 0.7872
Epoch 2/20
6570/7200 [==========================>...] - ETA: 0s - loss: 0.4978 - acc: 0.7992Epoch 2/20
 330/7200 [>.............................] - ETA: 3s - loss: 0.4336 - acc: 0.7939
 660/7200 [=>............................] - ETA: 3s - loss: 0.4388 - acc: 0.7894Epoch 2/20
1100/7200 [===>..........................] - ETA: 3s - loss: 0.4425 - acc: 0.7845
1060/7200 [===>..........................] - ETA: 3s - loss: 0.4344 - acc: 0.7943Epoch 2/20
7200/7200 [==============================] - 3s - loss: 0.4250 - acc: 0.7975     
Epoch 3/20
7200/7200 [==============================] - 3s - loss: 0.4267 - acc: 0.7993     
6640/7200 [==========================>...] - ETA: 0s - loss: 0.4240 - acc: 0.7974Epoch 3/20
6240/7200 [=========================>....] - ETA: 0s - loss: 0.4265 - acc: 0.7974
  80/7200 [..............................] - ETA: 5s - loss: 0.4717 - acc: 0.8125Epoch 3/20
  10/7200 [..............................] - ETA: 2s - loss: 0.3999 - acc: 0.9000
6970/7200 [============================>.] - ETA: 0s - loss: 0.4244 - acc: 0.7987Epoch 3/20
 690/7200 [=>............................] - ETA: 3s - loss: 0.4352 - acc: 0.8000
 420/7200 [>.............................] - ETA: 4s - loss: 0.4503 - acc: 0.7976Epoch 3/20
 450/7200 [>.............................] - ETA: 3s - loss: 0.4309 - acc: 0.8022
 790/7200 [==>...........................] - ETA: 3s - loss: 0.4397 - acc: 0.7911Epoch 3/20
1020/7200 [===>..........................] - ETA: 3s - loss: 0.4409 - acc: 0.7853
Epoch 3/20
1420/7200 [====>.........................] - ETA: 3s - loss: 0.4423 - acc: 0.7775
1090/7200 [===>..........................] - ETA: 3s - loss: 0.4501 - acc: 0.7771Epoch 3/20
6910/7200 [===========================>..] - ETA: 0s - loss: 0.4161 - acc: 0.8013
7000/7200 [============================>.] - ETA: 0s - loss: 0.4191 - acc: 0.7979Epoch 4/20
5940/7200 [=======================>......] - ETA: 0s - loss: 0.4244 - acc: 0.7931
6620/7200 [==========================>...] - ETA: 0s - loss: 0.4219 - acc: 0.7982Epoch 4/20
6120/7200 [========================>.....] - ETA: 0s - loss: 0.4234 - acc: 0.7940
6260/7200 [=========================>....] - ETA: 0s - loss: 0.4213 - acc: 0.7989Epoch 4/20
6380/7200 [=========================>....] - ETA: 0s - loss: 0.4205 - acc: 0.7998
 340/7200 [>.............................] - ETA: 3s - loss: 0.4381 - acc: 0.8147Epoch 4/20
6590/7200 [==========================>...] - ETA: 0s - loss: 0.4202 - acc: 0.7997
 450/7200 [>.............................] - ETA: 3s - loss: 0.4318 - acc: 0.8067Epoch 4/20
 390/7200 [>.............................] - ETA: 3s - loss: 0.4586 - acc: 0.7949
 640/7200 [=>............................] - ETA: 3s - loss: 0.4429 - acc: 0.8047Epoch 4/20
1050/7200 [===>..........................] - ETA: 3s - loss: 0.4400 - acc: 0.7914
 570/7200 [=>............................] - ETA: 3s - loss: 0.4676 - acc: 0.7982Epoch 4/20
 110/7200 [..............................] - ETA: 3s - loss: 0.4381 - acc: 0.8000
1250/7200 [====>.........................] - ETA: 3s - loss: 0.4409 - acc: 0.7904Epoch 4/20
7200/7200 [==============================] - 3s - loss: 0.4154 - acc: 0.8093     
Epoch 5/20
6850/7200 [===========================>..] - ETA: 0s - loss: 0.4176 - acc: 0.8055
6860/7200 [===========================>..] - ETA: 0s - loss: 0.4171 - acc: 0.8086Epoch 5/20
7060/7200 [============================>.] - ETA: 0s - loss: 0.4154 - acc: 0.8099
 190/7200 [..............................] - ETA: 4s - loss: 0.3924 - acc: 0.8105Epoch 5/20
6200/7200 [========================>.....] - ETA: 0s - loss: 0.4163 - acc: 0.8087

Epoch 5/20
 320/7200 [>.............................] - ETA: 4s - loss: 0.3776 - acc: 0.8281Epoch 5/20
7200/7200 [==============================] - 3s - loss: 0.4128 - acc: 0.8260     
 490/7200 [=>............................] - ETA: 4s - loss: 0.4055 - acc: 0.8143Epoch 5/20
 750/7200 [==>...........................] - ETA: 3s - loss: 0.3717 - acc: 0.8293
6980/7200 [============================>.] - ETA: 0s - loss: 0.4149 - acc: 0.8103Epoch 5/20
 150/7200 [..............................] - ETA: 4s - loss: 0.4206 - acc: 0.8067
 930/7200 [==>...........................] - ETA: 3s - loss: 0.4085 - acc: 0.8312Epoch 5/20
7200/7200 [==============================] - 4s - loss: 0.4121 - acc: 0.8299     

6770/7200 [===========================>..] - ETA: 0s - loss: 0.4081 - acc: 0.8374Epoch 6/20
Epoch 6/20
 110/7200 [..............................] - ETA: 3s - loss: 0.3680 - acc: 0.8727
Epoch 6/20
6060/7200 [========================>.....] - ETA: 0s - loss: 0.4135 - acc: 0.8302
7060/7200 [============================>.] - ETA: 0s - loss: 0.4077 - acc: 0.8368Epoch 6/20

 350/7200 [>.............................] - ETA: 3s - loss: 0.3946 - acc: 0.8429Epoch 6/20
 570/7200 [=>............................] - ETA: 3s - loss: 0.4086 - acc: 0.8368
6660/7200 [==========================>...] - ETA: 0s - loss: 0.4146 - acc: 0.8254Epoch 6/20
 820/7200 [==>...........................] - ETA: 3s - loss: 0.4227 - acc: 0.8171
Epoch 6/20
1020/7200 [===>..........................] - ETA: 3s - loss: 0.4064 - acc: 0.8333
Epoch 6/20
7200/7200 [==============================] - 4s - loss: 0.4107 - acc: 0.8292     
7200/7200 [==============================] - 4s - loss: 0.4098 - acc: 0.8318     Epoch 7/20
6220/7200 [========================>.....] - ETA: 0s - loss: 0.4175 - acc: 0.8251
6990/7200 [============================>.] - ETA: 0s - loss: 0.4108 - acc: 0.8288Epoch 7/20
 100/7200 [..............................] - ETA: 3s - loss: 0.4096 - acc: 0.8200
6910/7200 [===========================>..] - ETA: 0s - loss: 0.4066 - acc: 0.8352Epoch 7/20
7200/7200 [==============================] - 4s - loss: 0.4118 - acc: 0.8289     
6050/7200 [========================>.....] - ETA: 0s - loss: 0.4085 - acc: 0.8311
7040/7200 [============================>.] - ETA: 0s - loss: 0.4047 - acc: 0.8365Epoch 7/20
 310/7200 [>.............................] - ETA: 3s - loss: 0.3850 - acc: 0.8355Epoch 7/20
 150/7200 [..............................] - ETA: 5s - loss: 0.3484 - acc: 0.8533
Epoch 7/20
 820/7200 [==>...........................] - ETA: 3s - loss: 0.4010 - acc: 0.8207
 870/7200 [==>...........................] - ETA: 3s - loss: 0.4014 - acc: 0.8230Epoch 7/20
7200/7200 [==============================] - 4s - loss: 0.4098 - acc: 0.8312     
1580/7200 [=====>........................] - ETA: 2s - loss: 0.4065 - acc: 0.8228Epoch 7/20
6810/7200 [===========================>..] - ETA: 0s - loss: 0.4105 - acc: 0.8325
5610/7200 [======================>.......] - ETA: 0s - loss: 0.4047 - acc: 0.8332Epoch 8/20
  10/7200 [..............................] - ETA: 3s - loss: 0.2223 - acc: 0.9000

7060/7200 [============================>.] - ETA: 0s - loss: 0.4100 - acc: 0.8319Epoch 8/20
5760/7200 [=======================>......] - ETA: 0s - loss: 0.4051 - acc: 0.8325Epoch 8/20
7200/7200 [==============================] - 4s - loss: 0.4097 - acc: 0.8321     
6460/7200 [=========================>....] - ETA: 0s - loss: 0.4131 - acc: 0.8319Epoch 8/20
7200/7200 [==============================] - 4s - loss: 0.4024 - acc: 0.8386     
 450/7200 [>.............................] - ETA: 3s - loss: 0.3752 - acc: 0.8556Epoch 8/20
 430/7200 [>.............................] - ETA: 3s - loss: 0.3716 - acc: 0.8442
6130/7200 [========================>.....] - ETA: 0s - loss: 0.4065 - acc: 0.8323Epoch 8/20
 590/7200 [=>............................] - ETA: 3s - loss: 0.3754 - acc: 0.8610
1120/7200 [===>..........................] - ETA: 3s - loss: 0.3938 - acc: 0.8393Epoch 8/20
 900/7200 [==>...........................] - ETA: 3s - loss: 0.3869 - acc: 0.8567
Epoch 8/20
6990/7200 [============================>.] - ETA: 0s - loss: 0.4041 - acc: 0.8369
Epoch 9/20
7200/7200 [==============================] - 4s - loss: 0.4055 - acc: 0.8357     
5760/7200 [=======================>......] - ETA: 0s - loss: 0.4052 - acc: 0.8352Epoch 9/20
7160/7200 [============================>.] - ETA: 0s - loss: 0.4084 - acc: 0.8325
Epoch 9/20
6210/7200 [========================>.....] - ETA: 0s - loss: 0.4069 - acc: 0.8346
 330/7200 [>.............................] - ETA: 5s - loss: 0.5118 - acc: 0.7909Epoch 9/20
 440/7200 [>.............................] - ETA: 4s - loss: 0.4802 - acc: 0.8023
 240/7200 [>.............................] - ETA: 3s - loss: 0.4665 - acc: 0.8208Epoch 9/20
 920/7200 [==>...........................] - ETA: 3s - loss: 0.4135 - acc: 0.8293
Epoch 9/20
1370/7200 [====>.........................] - ETA: 3s - loss: 0.4108 - acc: 0.8328
1550/7200 [=====>........................] - ETA: 3s - loss: 0.4196 - acc: 0.8297Epoch 9/20
1200/7200 [====>.........................] - ETA: 3s - loss: 0.4226 - acc: 0.8233
1440/7200 [=====>........................] - ETA: 3s - loss: 0.4106 - acc: 0.8375Epoch 9/20
7200/7200 [==============================] - 3s - loss: 0.4052 - acc: 0.8354     
Epoch 10/20
 430/7200 [>.............................] - ETA: 4s - loss: 0.4199 - acc: 0.8326
5960/7200 [=======================>......] - ETA: 0s - loss: 0.4037 - acc: 0.8357Epoch 10/20

5900/7200 [=======================>......] - ETA: 0s - loss: 0.4043 - acc: 0.8358Epoch 10/20
  60/7200 [..............................] - ETA: 6s - loss: 0.4578 - acc: 0.7833
7200/7200 [==============================] - 3s - loss: 0.3996 - acc: 0.8393     Epoch 10/20
6770/7200 [===========================>..] - ETA: 0s - loss: 0.4082 - acc: 0.8325
 580/7200 [=>............................] - ETA: 4s - loss: 0.4044 - acc: 0.8345Epoch 10/20
 680/7200 [=>............................] - ETA: 4s - loss: 0.4092 - acc: 0.8279
 580/7200 [=>............................] - ETA: 3s - loss: 0.4135 - acc: 0.8379Epoch 10/20
1090/7200 [===>..........................] - ETA: 3s - loss: 0.3873 - acc: 0.8459
1110/7200 [===>..........................] - ETA: 3s - loss: 0.3889 - acc: 0.8505Epoch 10/20
7200/7200 [==============================] - 4s - loss: 0.4076 - acc: 0.8346     
Epoch 10/20
6740/7200 [===========================>..] - ETA: 0s - loss: 0.4036 - acc: 0.8362
Epoch 11/20
6990/7200 [============================>.] - ETA: 0s - loss: 0.4037 - acc: 0.8365
7120/7200 [============================>.] - ETA: 0s - loss: 0.4022 - acc: 0.8369Epoch 11/20
7200/7200 [==============================] - 4s - loss: 0.3984 - acc: 0.8382     

6010/7200 [========================>.....] - ETA: 0s - loss: 0.4051 - acc: 0.8361Epoch 11/20
  80/7200 [..............................] - ETA: 6s - loss: 0.4434 - acc: 0.8000 Epoch 11/20
6610/7200 [==========================>...] - ETA: 0s - loss: 0.4023 - acc: 0.8384
6210/7200 [========================>.....] - ETA: 0s - loss: 0.4012 - acc: 0.8374Epoch 11/20
7200/7200 [==============================] - 4s - loss: 0.4046 - acc: 0.8364     
 950/7200 [==>...........................] - ETA: 3s - loss: 0.3857 - acc: 0.8411Epoch 11/20
7130/7200 [============================>.] - ETA: 0s - loss: 0.4056 - acc: 0.8355
1190/7200 [===>..........................] - ETA: 3s - loss: 0.3802 - acc: 0.8395Epoch 11/20
7200/7200 [==============================] - 4s - loss: 0.4064 - acc: 0.8350     
  10/7200 [..............................] - ETA: 6s - loss: 0.2716 - acc: 0.9000Epoch 11/20
6040/7200 [========================>.....] - ETA: 0s - loss: 0.4007 - acc: 0.8382Epoch 13/20

Epoch 13/20
 350/7200 [>.............................] - ETA: 4s - loss: 0.4089 - acc: 0.8286
6000/7200 [========================>.....] - ETA: 0s - loss: 0.4015 - acc: 0.8372Epoch 13/20
 360/7200 [>.............................] - ETA: 4s - loss: 0.4113 - acc: 0.8389
 740/7200 [==>...........................] - ETA: 4s - loss: 0.4180 - acc: 0.8351Epoch 13/20
6910/7200 [===========================>..] - ETA: 0s - loss: 0.4032 - acc: 0.8359
Epoch 13/20
1600/7200 [=====>........................] - ETA: 3s - loss: 0.3940 - acc: 0.8419
1290/7200 [====>.........................] - ETA: 3s - loss: 0.4137 - acc: 0.8326Epoch 13/20
5350/7200 [=====================>........] - ETA: 1s - loss: 0.3960 - acc: 0.8394
6770/7200 [===========================>..] - ETA: 0s - loss: 0.4024 - acc: 0.8366Epoch 15/20
7200/7200 [==============================] - 4s - loss: 0.3950 - acc: 0.8382     
Epoch 15/20
7200/7200 [==============================] - 4s - loss: 0.4015 - acc: 0.8372     
6110/7200 [========================>.....] - ETA: 0s - loss: 0.4004 - acc: 0.8401Epoch 15/20
6710/7200 [==========================>...] - ETA: 0s - loss: 0.3978 - acc: 0.8399
 530/7200 [=>............................] - ETA: 4s - loss: 0.3898 - acc: 0.8302Epoch 15/20
6350/7200 [=========================>....] - ETA: 0s - loss: 0.3997 - acc: 0.8402
5960/7200 [=======================>......] - ETA: 0s - loss: 0.3989 - acc: 0.8384Epoch 15/20
 560/7200 [=>............................] - ETA: 4s - loss: 0.3755 - acc: 0.8411
 630/7200 [=>............................] - ETA: 4s - loss: 0.3916 - acc: 0.8365Epoch 15/20
1040/7200 [===>..........................] - ETA: 4s - loss: 0.4027 - acc: 0.8279
Epoch 15/20
 430/7200 [>.............................] - ETA: 5s - loss: 0.4077 - acc: 0.8395
Epoch 15/20
6450/7200 [=========================>....] - ETA: 0s - loss: 0.4038 - acc: 0.8366
5230/7200 [====================>.........] - ETA: 1s - loss: 0.4001 - acc: 0.8377Epoch 16/20
7160/7200 [============================>.] - ETA: 0s - loss: 0.4027 - acc: 0.8363
6430/7200 [=========================>....] - ETA: 0s - loss: 0.4046 - acc: 0.8364Epoch 16/20
6120/7200 [========================>.....] - ETA: 0s - loss: 0.4037 - acc: 0.8364

Epoch 16/20
Epoch 16/20
6190/7200 [========================>.....] - ETA: 0s - loss: 0.4045 - acc: 0.8346
6900/7200 [===========================>..] - ETA: 0s - loss: 0.4014 - acc: 0.8378Epoch 16/20
7200/7200 [==============================] - 4s - loss: 0.4005 - acc: 0.8382     
 750/7200 [==>...........................] - ETA: 4s - loss: 0.4081 - acc: 0.8360Epoch 16/20
1720/7200 [======>.......................] - ETA: 3s - loss: 0.4145 - acc: 0.8291
 400/7200 [>.............................] - ETA: 4s - loss: 0.4244 - acc: 0.8275Epoch 16/20
 680/7200 [=>............................] - ETA: 4s - loss: 0.4109 - acc: 0.8338
Epoch 16/20
6500/7200 [==========================>...] - ETA: 0s - loss: 0.4013 - acc: 0.8392
 570/7200 [=>............................] - ETA: 3s - loss: 0.4218 - acc: 0.8263Epoch 17/20
7200/7200 [==============================] - 4s - loss: 0.4015 - acc: 0.8382     
6040/7200 [========================>.....] - ETA: 0s - loss: 0.4025 - acc: 0.8379
7130/7200 [============================>.] - ETA: 0s - loss: 0.3994 - acc: 0.8374Epoch 17/20
6080/7200 [========================>.....] - ETA: 0s - loss: 0.4085 - acc: 0.8347Epoch 17/20
 120/7200 [..............................] - ETA: 3s - loss: 0.3848 - acc: 0.8583
6650/7200 [==========================>...] - ETA: 0s - loss: 0.4001 - acc: 0.8395Epoch 17/20
6780/7200 [===========================>..] - ETA: 0s - loss: 0.4034 - acc: 0.8364
Epoch 17/20
 440/7200 [>.............................] - ETA: 4s - loss: 0.4192 - acc: 0.8227
 970/7200 [===>..........................] - ETA: 4s - loss: 0.4100 - acc: 0.8340Epoch 17/20
7200/7200 [==============================] - 4s - loss: 0.4004 - acc: 0.8369     
Epoch 17/20
7200/7200 [==============================] - 4s - loss: 0.3970 - acc: 0.8397     
5770/7200 [=======================>......] - ETA: 0s - loss: 0.4039 - acc: 0.8348
7200/7200 [==============================] - 4s - loss: 0.3994 - acc: 0.8383     
7200/7200 [==============================] - 4s - loss: 0.3972 - acc: 0.8371     

7200/7200 [==============================] - 4s - loss: 0.3975 - acc: 0.8378     
7200/7200 [==============================] - 4s - loss: 0.3978 - acc: 0.8387     
7200/7200 [==============================] - 4s - loss: 0.3995 - acc: 0.8368     
730/800 [==========================>...] - ETA: 0sEpoch 1/20
Epoch 1/20
7200/7200 [==============================] - 2s - loss: 0.4881 - acc: 0.7960     
Epoch 2/20
7200/7200 [==============================] - 2s - loss: 0.4905 - acc: 0.7957     
Epoch 2/20
7200/7200 [==============================] - 2s - loss: 0.4244 - acc: 0.7968     
Epoch 3/20
7200/7200 [==============================] - 2s - loss: 0.4265 - acc: 0.7965     
Epoch 3/20
7200/7200 [==============================] - 2s - loss: 0.4188 - acc: 0.7983     
Epoch 4/20
7200/7200 [==============================] - 2s - loss: 0.4215 - acc: 0.7965     
Epoch 4/20
7200/7200 [==============================] - 2s - loss: 0.4143 - acc: 0.8257     
Epoch 5/20
7200/7200 [==============================] - 2s - loss: 0.4173 - acc: 0.8107     
Epoch 5/20
7200/7200 [==============================] - 2s - loss: 0.4114 - acc: 0.8317     
Epoch 6/20
7200/7200 [==============================] - 2s - loss: 0.4148 - acc: 0.8246     
Epoch 6/20
7200/7200 [==============================] - 1s - loss: 0.4101 - acc: 0.8329     
Epoch 7/20
7200/7200 [==============================] - 1s - loss: 0.4135 - acc: 0.8296     
Epoch 7/20
7200/7200 [==============================] - 1s - loss: 0.4082 - acc: 0.8346     
Epoch 8/20
7200/7200 [==============================] - 1s - loss: 0.4117 - acc: 0.8311     
Epoch 8/20
7200/7200 [==============================] - 1s - loss: 0.4071 - acc: 0.8367     
Epoch 9/20
7200/7200 [==============================] - 1s - loss: 0.4104 - acc: 0.8315     
Epoch 9/20
7200/7200 [==============================] - 1s - loss: 0.4060 - acc: 0.8364     
Epoch 10/20
7200/7200 [==============================] - 1s - loss: 0.4095 - acc: 0.8319     
Epoch 10/20
7200/7200 [==============================] - 1s - loss: 0.4048 - acc: 0.8369     
Epoch 11/20
7200/7200 [==============================] - 1s - loss: 0.4081 - acc: 0.8324     
Epoch 11/20
7200/7200 [==============================] - 1s - loss: 0.4038 - acc: 0.8365     
Epoch 12/20
7200/7200 [==============================] - 1s - loss: 0.4072 - acc: 0.8329     
Epoch 12/20
7200/7200 [==============================] - 1s - loss: 0.4031 - acc: 0.8374     
Epoch 13/20
7200/7200 [==============================] - 1s - loss: 0.4061 - acc: 0.8336     
Epoch 13/20
7200/7200 [==============================] - 1s - loss: 0.4025 - acc: 0.8378     
Epoch 14/20
7200/7200 [==============================] - 1s - loss: 0.4058 - acc: 0.8329     
Epoch 14/20
7200/7200 [==============================] - 1s - loss: 0.4015 - acc: 0.8381     
Epoch 15/20
7200/7200 [==============================] - 1s - loss: 0.4048 - acc: 0.8336     
Epoch 15/20
7200/7200 [==============================] - 2s - loss: 0.4007 - acc: 0.8396     
Epoch 16/20
7200/7200 [==============================] - 2s - loss: 0.4041 - acc: 0.8351     
Epoch 16/20
7200/7200 [==============================] - 1s - loss: 0.4005 - acc: 0.8382     
Epoch 17/20
7200/7200 [==============================] - 1s - loss: 0.4038 - acc: 0.8344     
Epoch 17/20
7200/7200 [==============================] - 1s - loss: 0.4002 - acc: 0.8376     
Epoch 18/20
7200/7200 [==============================] - 1s - loss: 0.4034 - acc: 0.8353     
Epoch 18/20
7200/7200 [==============================] - 1s - loss: 0.3994 - acc: 0.8381     
5980/7200 [=======================>......] - ETA: 0s - loss: 0.4016 - acc: 0.8363Epoch 19/20
7200/7200 [==============================] - 1s - loss: 0.4025 - acc: 0.8346     
Epoch 19/20
7200/7200 [==============================] - 1s - loss: 0.3992 - acc: 0.8390     
Epoch 20/20
7200/7200 [==============================] - 1s - loss: 0.4024 - acc: 0.8349     
Epoch 20/20
7200/7200 [==============================] - 1s - loss: 0.3979 - acc: 0.8381     
7200/7200 [==============================] - 1s - loss: 0.4015 - acc: 0.8349     
 10/800 [..............................] - ETA: 2s
In [25]:
accuracies
Out[25]:
array([ 0.82999999,  0.82875   ,  0.845     ,  0.83624999,  0.83499999,
        0.83874999,  0.83875   ,  0.83249999,  0.82499999,  0.85624999])
In [26]:
accuracies.mean()
Out[26]:
0.83662499487400055
In [27]:
accuracies.std()
Out[27]:
0.0085156105222930732
By at
Labels: , ,

No comments :

Post a Comment

Subscribe to: Post Comments ( Atom )