Tensorflow--邏輯迴歸

#coding=utf-8
import tensorflow as tf
import numpy as np
num_point = 100
vectors_set = []
#train data
for i in range(num_point):
	x1 = np.random.normal(0.0,1)
	y1 = 1 if x1*0.3+0.1 +np.random.normal(0.0,0.03)>0 else 0
	vectors_set.append([x1,y1])
x_data = [v[0] for v in vectors_set]
y_data = [v[1] for v in vectors_set]


w = tf.Variable(tf.random_uniform([1],-1.0,1.0))
b = tf.Variable(tf.zeros([1]))

y = tf.sigmoid(w*x_data+b)

one = tf.ones(y.get_shape(),dtype = tf.float32)
#交叉熵損失函數
loss = -tf.reduce_mean(y_data*tf.log(y)+(one-y_data)*tf.log(one-y))
#梯度下降學習算法
train = tf.train.GradientDescentOptimizer(0.5).minimize(loss)
with tf.Session() as sess:
	sess.run(tf.initialize_all_variables())
	th = tf.ones_like(one,dtype = tf.float32)*0.5
	#tf.cast(x,dtype)將x的數據格式轉化爲dtype
	#評估
	correct_prediction = tf.equal(tf.cast(y_data,tf.int32),tf.cast(tf.greater(y,th),tf.int32))
	accuracy = tf.reduce_mean(tf.cast(correct_prediction,tf.float32))
	for i in range(200):
		sess.run(train)
		if i%20==0:
			print ("accuracy",sess.run(accuracy))
			print ("loss",sess.run(loss))
			#print ('y',y_data)
			#print ("Y_predict",sess.run(y))