import tensorflow as tf
import numpy as np
import matplotlib.pyplot as plt
x_data=np.linspace(-0.5,0.5,200)[:,np.newaxis]
noise=np.random.normal(0,0.02,x_data.shape)
y_data=np.square(x_data)+noise
x=tf.placeholder(tf.float32,[None,1])
y=tf.placeholder(tf.float32,[None,1])
wei1=tf.Variable(tf.random_normal([1,10]))
bias1=tf.Variable(tf.zeros([1,10]))
wx1=tf.matmul(x,wei1)+bias1
L1=tf.nn.tanh(wx1)#激活函數

wei2=tf.Variable(tf.random_normal([10,1]))#入10出1
bias2=tf.Variable(tf.zeros([1,1]))
wx2=tf.matmul(L1,wei2)+bias2
prediction=tf.nn.tanh(wx2)

#二次代價
loss=tf.reduce_mean(tf.square(y-prediction))
train_step=tf.train.GradientDescentOptimizer(0.1).minimize(loss)

ini=tf.global_variables_initializer()

with tf.Session() as sess:
    sess.run(ini)
    for _ in range(2000):
        sess.run(train_step,feed_dict={x:x_data,y:y_data})
    pre_value=sess.run(prediction,feed_dict={x:x_data})
    plt.figure()
    plt.scatter(x_data,y_data)
    plt.plot(x_data,pre_value,'g-',lw=4)
    plt.show()