In [1]:

    

import tensorflow as tf


a = tf.constant([15.0,16.0,17.0,12.0,11.0,10.0,9.0],name ='x')
b = tf.constant([7.0,1.0,2.0,5.0,4.0,9.0,6.0],name='y')
training_epoch = 500
learning_rate = 0.005

t1=0.0
t0=0.0
##---------------------------finding the value of h0------------


h0=tf.divide(1.0,tf.add(1.0,tf.exp(-tf.add(t0,tf.reduce_sum(tf.multiply(t1,a))))))
   
model=tf.global_variables_initializer()
sess = tf.Session()
print("h0 values:",sess.run(h0))
result=tf.cond(h0<0.5,lambda:0.00,lambda:1.00)
print("after rounding off, h0 value:")
print(sess.run(result))

############################################################
##-----------------------cost function-----------
c1 = tf.multiply(tf.subtract(1.0,b),tf.log(tf.subtract(1.0,result)))
c2 = tf.multiply(-b,tf.log(result))
cost = tf.reduce_sum(tf.subtract(c2,c1))
print("cost function:",sess.run(cost))

####----------------finding the value of theta 1 and theta 0-----------

for epoch in range(training_epoch):
    
    #if error is greater than threshold value then stop the iteration
    s1=tf.reduce_sum(tf.multiply(tf.subtract(tf.add(tf.multiply(a,t1),t0),b),a))
    s0=tf.reduce_sum(tf.subtract(tf.add(tf.multiply(a,t1),t0),b))
    tn1 = tf.subtract(t1,tf.multiply(learning_rate,tf.divide(s1,tf.cast(tf.size(a),tf.float32))))
    tn0=  tf.subtract(t0,tf.multiply(learning_rate,tf.divide(s0,tf.cast(tf.size(a),tf.float32))))
    t1=tn1
    t0=tn0
    model=tf.global_variables_initializer()
sess = tf.Session()
print("After 500 epoch values:")
print("theta 1 : ",sess.run(tn1))
print("theta 0 : ",sess.run(tn0))


    

    




h0 values: 0.5
after rounding off, h0 value:
1.0
cost function: nan
After 500 epoch values:
theta 1 :  0.226671
theta 0 :  1.41823

In [ ]: