tensorflow处理简单线性回归

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

num_points = 100
set =[]
for i in range(num_points):
    x1 = np.random.rand()
    y1 = x1  + 0.3 + np.random.normal(0.0, 0.03)
    set.append([x1,y1])
# 生成一些样本
x_data = [v[0] for v in set]
y_data = [v[1] for v in 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
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')

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

# 初始化的W和b是多少
print ("W =", sess.run(W), "b =",  "lossess.run(b)s =", sess.run(loss))
# 执行20次训练
for step in range(20):
    sess.run(train)
    # 输出训练好的W和b
    print ("W =", sess.run(W), "b =",  "lossess.run(b)s =", sess.run(loss))

W = [0.9896785] b = lossess.run(b)s = 0.00086038024
W = [0.9905102] b = lossess.run(b)s = 0.0008586294
W = [0.9912857] b = lossess.run(b)s = 0.00085710763
W = [0.9920086] b = lossess.run(b)s = 0.0008557848
W = [0.99268264] b = lossess.run(b)s = 0.0008546348
W = [0.99331105] b = lossess.run(b)s = 0.0008536352
W = [0.99389696] b = lossess.run(b)s = 0.00085276604
W = [0.9944432] b = lossess.run(b)s = 0.0008520109
W = [0.99495244] b = lossess.run(b)s = 0.0008513546
W = [0.99542725] b = lossess.run(b)s = 0.000850784
W = [0.99586993] b = lossess.run(b)s = 0.00085028785
W = [0.99628264] b = lossess.run(b)s = 0.00084985676
W = [0.99666744] b = lossess.run(b)s = 0.00084948225
W = [0.9970262] b = lossess.run(b)s = 0.0008491559
W = [0.99736065] b = lossess.run(b)s = 0.0008488733
W = [0.9976725] b = lossess.run(b)s = 0.000848627
W = [0.9979632] b = lossess.run(b)s = 0.000848413
W = [0.9982343] b = lossess.run(b)s = 0.00084822683
W = [0.998487] b = lossess.run(b)s = 0.0008480652
W = [0.9987226] b = lossess.run(b)s = 0.00084792456
W = [0.99894226] b = lossess.run(b)s = 0.00084780285

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