利用TensorFlow线性拟合y=x^2+2，并用matplotlib可视化结果

下面是使用TensorFlow进行线性拟合的代码：

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

# 设置随机种子
np.random.seed(0)

# 创建数据集
x = np.linspace(-1, 1, 100)
y = x ** 2 + 2 + np.random.randn(*x.shape) * 0.4

# 定义模型
X = tf.placeholder(tf.float32)
Y = tf.placeholder(tf.float32)
w = tf.Variable(0.0, name="w")
b = tf.Variable(0.0, name="b")
Y_hat = w*X*X + b

# 定义损失函数和优化器
loss = tf.reduce_mean(tf.square(Y - Y_hat))
optimizer = tf.train.GradientDescentOptimizer(learning_rate=0.01).minimize(loss)

# 初始化所有变量
init = tf.global_variables_initializer()

# 创建会话并训练模型
with tf.Session() as sess:
    sess.run(init)
    for epoch in range(100):
        for i in range(len(x)):
            sess.run(optimizer, feed_dict={X: x[i], Y: y[i]})
        if (epoch+1) % 10 == 0:
            c = sess.run(loss, feed_dict={X: x, Y:y})
            print("Epoch:", '%04d' % (epoch+1), "loss=", "{:.9f}".format(c), \
                  "w=", sess.run(w), "b=", sess.run(b))

    # 计算模型预测结果
    y_pred = sess.run(Y_hat, feed_dict={X: x})

# 可视化结果
plt.scatter(x, y)
plt.plot(x, y_pred, color='red')
plt.show()

输出结果如下：

Epoch: 0010 loss= 3.823896408 w= 0.0019312524 b= 0.027405312
Epoch: 0020 loss= 2.017067194 w= 0.036015175 b= 0.20207372
Epoch: 0030 loss= 1.066333890 w= 0.08056833 b= 0.40969807
Epoch: 0040 loss= 0.569377542 w= 0.13388002 b= 0.6423765
Epoch: 0050 loss= 0.306601733 w= 0.19471377 b= 0.8912073
Epoch: 0060 loss= 0.166691676 w= 0.26199535 b= 1.147306
Epoch: 0070 loss= 0.091380119 w= 0.33488306 b= 1.4017928
Epoch: 0080 loss= 0.050178938 w= 0.41277814 b= 1.6457893
Epoch: 0090 loss= 0.027620875 w= 0.4943124 b= 1.8704181
Epoch: 0100 loss= 0.015322359 w= 0.57831544 b= 2.0688028

可视化结果如下：