In [3]:

    

import tensorflow as tf


    

In [4]:

    

print(tf.__version__)


    

    




1.2.0

In [6]:

    

import keras as kr


    

In [7]:

    

print(kr.__version__)


    

    




2.1.2

In [8]:

    

from keras.models import Sequential
from keras.layers import Dense


    

In [9]:

    

import numpy as np


    

In [17]:

    

seed = 0
np.random.seed(seed)
tf.set_random_seed(seed)


    

In [18]:

    

Data_set = np.loadtxt("./datasets/ThoraricSurgery.csv", delimiter=",")


    

In [19]:

    

X = Data_set[:,0:17]


    

In [22]:

    

Y = Data_set[:,17]


    

In [23]:

    

print(Y.shape)


    

    




(470,)

In [24]:

    

model = Sequential()
model.add(Dense(30, input_dim = 17, activation='relu'))
model.add(Dense(1, activation='sigmoid'))


    

In [25]:

    

model.compile(loss='mean_squared_error', optimizer='adam',metrics=['accuracy'])


    

In [26]:

    

model.fit(X, Y, epochs=30, batch_size=10)


    

    




Epoch 1/30
470/470 [==============================] - 1s 1ms/step - loss: 0.6612 - acc: 0.3149
Epoch 2/30
470/470 [==============================] - 0s 322us/step - loss: 0.1488 - acc: 0.8511
Epoch 3/30
470/470 [==============================] - 0s 318us/step - loss: 0.1488 - acc: 0.8511
Epoch 4/30
470/470 [==============================] - 0s 320us/step - loss: 0.1488 - acc: 0.8511
Epoch 5/30
470/470 [==============================] - 0s 317us/step - loss: 0.1488 - acc: 0.8511
Epoch 6/30
470/470 [==============================] - 0s 318us/step - loss: 0.1487 - acc: 0.8511
Epoch 7/30
470/470 [==============================] - 0s 320us/step - loss: 0.1487 - acc: 0.8511
Epoch 8/30
470/470 [==============================] - 0s 319us/step - loss: 0.1487 - acc: 0.8511
Epoch 9/30
470/470 [==============================] - 0s 326us/step - loss: 0.1487 - acc: 0.8511
Epoch 10/30
470/470 [==============================] - 0s 322us/step - loss: 0.1486 - acc: 0.8511
Epoch 11/30
470/470 [==============================] - 0s 316us/step - loss: 0.1499 - acc: 0.8447
Epoch 12/30
470/470 [==============================] - 0s 334us/step - loss: 0.1486 - acc: 0.8511
Epoch 13/30
470/470 [==============================] - 0s 327us/step - loss: 0.1485 - acc: 0.8511
Epoch 14/30
470/470 [==============================] - 0s 335us/step - loss: 0.1483 - acc: 0.8511
Epoch 15/30
470/470 [==============================] - 0s 349us/step - loss: 0.1485 - acc: 0.8511
Epoch 16/30
470/470 [==============================] - 0s 345us/step - loss: 0.1490 - acc: 0.8447
Epoch 17/30
470/470 [==============================] - 0s 418us/step - loss: 0.1478 - acc: 0.8489
Epoch 18/30
470/470 [==============================] - 0s 351us/step - loss: 0.1482 - acc: 0.8468
Epoch 19/30
470/470 [==============================] - 0s 357us/step - loss: 0.1476 - acc: 0.8511
Epoch 20/30
470/470 [==============================] - 0s 336us/step - loss: 0.1480 - acc: 0.8511
Epoch 21/30
470/470 [==============================] - 0s 313us/step - loss: 0.1474 - acc: 0.8511
Epoch 22/30
470/470 [==============================] - 0s 398us/step - loss: 0.1469 - acc: 0.8511
Epoch 23/30
470/470 [==============================] - 0s 444us/step - loss: 0.1466 - acc: 0.8511
Epoch 24/30
470/470 [==============================] - 0s 414us/step - loss: 0.1475 - acc: 0.8489
Epoch 25/30
470/470 [==============================] - 0s 332us/step - loss: 0.1470 - acc: 0.8511
Epoch 26/30
470/470 [==============================] - 0s 363us/step - loss: 0.1466 - acc: 0.8511
Epoch 27/30
470/470 [==============================] - 0s 376us/step - loss: 0.1472 - acc: 0.8511
Epoch 28/30
470/470 [==============================] - 0s 337us/step - loss: 0.1471 - acc: 0.8511
Epoch 29/30
470/470 [==============================] - 0s 322us/step - loss: 0.1470 - acc: 0.8489
Epoch 30/30
470/470 [==============================] - 0s 399us/step - loss: 0.1461 - acc: 0.8532






    
Out[26]:





<keras.callbacks.History at 0x11b7d0c88>

In [27]:

    

print("\n Accuracy: %.4f" % (model.evaluate(X, Y)[1]))


    

    




470/470 [==============================] - 0s 91us/step

 Accuracy: 0.8511

In [ ]: