（三）Tensorflow的多项式回归模型

多项式回归，代码展示，数据展示，效果展示，tensorboard定义图的展示

代码展示：

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


def polynomial_regression():
    #生成数据
    n_object = 100
    xs = np.linspace(-3,3,n_object)
    ys = np.sin(xs) + np.random.uniform(-0.5,0.5,n_object)
    plt.figure(figsize=[8,6])
    plt.scatter(xs,ys)
    plt.show()
    #给数据建立placeholder
    X = tf.placeholder(tf.float32,name="X")
    Y = tf.placeholder(tf.float32,name="Y")
    #初始化权重和偏置
    weight = tf.Variable(tf.random_normal([1]),name="weight")
    bias = tf.Variable(tf.random_normal([1]),name="bias")
    y_pred = tf.add(tf.multiply(weight,X),bias)
    #添加多项式
    weight2 = tf.Variable(tf.random_normal([1]), name="weight2")
    y_pred = tf.add(y_pred,tf.multiply(tf.pow(X,2),weight2))
    weight3 = tf.Variable(tf.random_normal([1]), name="weight3")
    y_pred = tf.add(y_pred,tf.multiply(tf.pow(X,3),weight3))
    #定义loss和优化器
    loss = tf.reduce_sum(tf.square(Y-y_pred),name="loss")/xs.shape[0]  #直接reduce_mean的话数值直接溢出了!
    train = tf.train.GradientDescentOptimizer(learning_rate=1e-2).minimize(loss)
    #初始化变量
    init = tf.global_variables_initializer()
    #通过session运行op
    with tf.Session() as sess:
        sess.run(init)
        writer = tf.summary.FileWriter("./logs",sess.graph)
        for i in range(1000):
            loss_total = 0
            for x_input,y_input in zip(xs,ys):
                _,l = sess.run([train,loss],feed_dict={X:x_input,Y:y_input})
                loss_total+=l
            if i%50==0:
                print("Iterator:{},loss:{}".format(i, loss_total/xs.shape[0]))
        writer.close()
        w1,w2,w3, b = sess.run([weight, weight2,weight3,bias])
        print("最终的  W1:{},W2:{},W3:{},b:{}".format(w1,w2,w3, b))
    plt.figure(figsize=[10,8])
    plt.scatter(xs,ys,label="real_data",c="r")
    plt.plot(xs,w1[0]*xs+w2[0]*np.power(xs,2)+w3[0]*np.power(xs,3),label="pred_data",c="b")
    plt.legend()
    plt.show()

数据展示：

效果展示：

tensorboard图展示：

有问题可以随时说明！！