import tensorflow as tf
import numpy as np
import matplotlib.pyplot as plt
x_data = np.linspace(-0.5, 0.5, 200)[:, np.newaxis] # 使得维度为[200, 1]
noise = np.random.normal(0, 0.02, x_data.shape) # 维度为[200, 1]的正太分布
y_data = np.square(x_data) + noise
# 定义两个占位符placeholder
x = tf.placeholder(tf.float32, [None, 1])
y = tf.placeholder(tf.float32, [None, 1])
# 定义神经网络的中间层
W1 = tf.Variable(tf.random_normal([1, 10])) # 10个神经元
b1 = tf.Variable(tf.zeros([1, 10]))
W1_plus_b1 = tf.matmul(x, W1) + b1
output1 = tf.nn.tanh(W1_plus_b1) # 激活函数tanh()
# 定义神经网络的输出层
W2 = tf.Variable(tf.random_normal([10, 1]))
b2 = tf.Variable(tf.zeros([1, 1]))
W2_plus_b2 = tf.matmul(output1, W2) + b2
prediction = tf.nn.tanh(W2_plus_b2) # 激活函数
# 二次代价函数
loss = tf.reduce_mean(tf.square(y - prediction))
train_step = tf.train.GradientDescentOptimizer(0.1).minimize(loss)
with tf.Session() as sess:
# 变量的初始化
sess.run(tf.global_variables_initializer())
# 训练2000次
for i in range(2000):
sess.run(train_step, feed_dict={x: x_data, y: y_data})
if i % 100 == 0:
print("step: "+str(i))
# 获取预测值
prediction_value = sess.run(prediction, feed_dict={x: x_data})
plt.figure()
plt.xlabel('x_data')
plt.ylabel('y_data')
plt.scatter(x_data, y_data)
plt.plot(x_data, prediction_value, 'r-', lw=5)
plt.show()
运行结果图如下: