# -*- coding:utf-8 -*-
#功能： 使用tensorflow實(shí)現(xiàn)一個(gè)簡(jiǎn)單的邏輯回歸
import tensorflow as tf
import numpy as np
import matplotlib.pyplot as plt
 
#創(chuàng)建占位符
X=tf.placeholder(tf.float32)
Y=tf.placeholder(tf.float32)
 
#創(chuàng)建變量
#tf.random_normal([1])返回一個(gè)符合正太分布的隨機(jī)數(shù)
w=tf.Variable(tf.random_normal([1],name='weight'))
b=tf.Variable(tf.random_normal([1],name='bias'))
 
y_predict=tf.sigmoid(tf.add(tf.mul(X,w),b))
num_samples=400
cost=tf.reduce_sum(tf.pow(y_predict-Y,2.0))/num_samples
 
#學(xué)習(xí)率
lr=0.01
optimizer=tf.train.AdamOptimizer().minimize(cost)
 
#創(chuàng)建session 并初始化所有變量
num_epoch=500
cost_accum=[]
cost_prev=0
#np.linspace（）創(chuàng)建agiel等差數(shù)組，元素個(gè)素為num_samples
xs=np.linspace(-5,5,num_samples)
ys=np.sin(xs)+np.random.normal(0,0.01,num_samples)
 
with tf.Session() as sess:
  #初始化所有變量
  sess.run(tf.initialize_all_variables())
  #開(kāi)始訓(xùn)練
  for epoch in range(num_epoch):
    for x,y in zip(xs,ys):
      sess.run(optimizer,feed_dict={X:x,Y:y})
    train_cost=sess.run(cost,feed_dict={X:x,Y:y})
    cost_accum.append(train_cost)
    print "train_cost is:",str(train_cost)
 
    #當(dāng)誤差小于10-6時(shí) 終止訓(xùn)練
    if np.abs(cost_prev-train_cost)<1e-6:
      break
    #保存最終的誤差
    cost_prev=train_cost
#畫(huà)圖 畫(huà)出每一輪訓(xùn)練所有樣本之后的誤差
plt.plot(range(len(cost_accum)),cost_accum,'r')
plt.title('Logic Regression Cost Curve')
plt.xlabel('epoch')
plt.ylabel('cost')
plt.show()