tensorflow入门第二篇：线性回归

利用TensorFlow建立线性回归模型：

代码如下：

import numpy as np
import tensorflow as tf
import matplotlib.pyplot as plt

# 随机生成1000个点，围绕在y=0.1x+0.3直线周围
num_points = 1000
vectors_set = []
for i in range(num_points):
    x1 = np.random.normal(0.0,0.55)
    y1 = x1*0.1 + 0.3 + np.random.normal(0.0,0.03)
    vectors_set.append([x1,y1])
    
# 生成样本
x_data = [v[0] for v in vectors_set]
x_data = [v[1] for v in vectors_set]

plt.scatter(x_data,y_data,c='r')
plt.show()

# 生成1维的W矩阵，取值是[-1,1]之间的随机数
w  = tf.Variable(tf.random_uniform([1],-1.0,1.0),name="W")

# 生成1维的b矩阵，初始值是0
b = tf.Variable(tf.zeros([1]),name="b")

#经过计算得出 预估值
y = w * x_data + b

# 以预估值y和实际值y_data之间的均方误差作为损失
loss = tf.reduce_mean(tf.square(y-y_data),name="loss")

# 采用梯度下降法来优化参数
optimizer = tf.train.GradientDescentOptimizer(0.5)

# 训练的过程就是最小化这个误差值
train = optimizer.minimize(loss,name="train")

sess = tf.Session()
init = tf.global_variables_initializer()
sess.run(init)

print("w=",sess.run(w),"b=",sess.run(b),"loss=",sess.run(loss))

#执行20次训练
for step in range(20):
    sess.run(train)
    print("w=",sess.run(w),"b=",sess.run(b),"loss=",sess.run(loss))

plt.scatter(x_data,y_data,c='r')
plt.plot(x_data,sess.run(w)*x_data+sess.run(b))
plt.show()

随机生成的样本数：

训练结果：

w= [-0.93157125] b= [0.] loss= 0.44114137
w= [-0.59994555] b= [0.3291781] loss= 0.15344466
w= [-0.3816231] b= [0.31994748] loss= 0.072886065
w= [-0.23160376] b= [0.3138706] loss= 0.034854285
w= [-0.12852593] b= [0.30969492] loss= 0.0168994
w= [-0.05770151] b= [0.3068258] loss= 0.008422874
w= [-0.0090383] b= [0.30485442] loss= 0.004421088
w= [0.02439804] b= [0.3034999] loss= 0.0025318374
w= [0.04737203] b= [0.30256924] loss= 0.0016399197
w= [0.06315739] b= [0.30192977] loss= 0.0012188433
w= [0.07400345] b= [0.3014904] loss= 0.0010200524
w= [0.08145574] b= [0.3011885] loss= 0.0009262026
w= [0.08657619] b= [0.30098107] loss= 0.00088189606
w= [0.09009444] b= [0.30083853] loss= 0.00086097885
w= [0.09251181] b= [0.30074063] loss= 0.00085110386
w= [0.09417278] b= [0.30067334] loss= 0.0008464418
w= [0.09531403] b= [0.30062708] loss= 0.000844241
w= [0.09609817] b= [0.30059534] loss= 0.0008432018
w= [0.09663695] b= [0.3005735] loss= 0.0008427114
w= [0.09700715] b= [0.3005585] loss= 0.0008424798
w= [0.09726151] b= [0.3005482] loss= 0.0008423704

可视化结果如下：