TensorFlow实现简单神经网络分类问题

import tensorflow as tf
import matplotlib.pyplot as plt
from numpy.random import RandomState
# Numpy 是一个科学计算的工具箱的，这里通过Numpy生成模拟数据集
 
# 定义训练数据batch的大小
batch_size = 8
 
w1 = tf.Variable(tf.random_normal([2, 3], stddev=1, seed=1))
w2 = tf.Variable(tf.random_normal([3, 1], stddev=1, seed=1))
 
x = tf.placeholder(tf.float32, shape=(None, 2), name='x-input')
y_ = tf.placeholder(tf.float32, shape=(None, 1), name='y-input')
 
# 定义神经网络前向传播过程
a = tf.matmul(x, w1)
y = tf.matmul(a, w2)
 
# 定义损失函数和反向传播算法
cross_entropy = -tf.reduce_mean(
    y_ * tf.log(tf.clip_by_value(y, 1e-10, 1.0)))
train_step = tf.train.AdamOptimizer(0.002).minimize(cross_entropy)
# 采用梯度下降优化器进行优化
# train_step = tf.train.GradientDescentOptimizer(0.05).minimize(cross_entropy)
 
 
# 通过随机数生成一个模拟数据集
rdm = RandomState(1)
dataset_size = 128
X = rdm.rand(dataset_size, 2)
 
# 产生样本标签，这里所有x1+x2<1的样例都被认为是正样本
Y = [[int(x1 + x2 < 1)] for (x1, x2) in X]
 
# 创建一个会话来运行tensorflow程序
with tf.Session() as sess:
    init_op = tf.initialize_all_variables()
    sess.run(init_op)
    print(sess.run(w1))
    print(sess.run(w2))
 
    STEPS = 5000
 
    plt.axis([0, 5000, 0, 0.08])
    plt.ion()
    x11 = [0, 0]
    y11 = [1, 1]
 
    for i in range(STEPS):
        start = (i * batch_size) % dataset_size
        end = min(start + batch_size, dataset_size)
        # 通过选取的样本训练神经网络并更新参数
        sess.run(train_step, feed_dict={x: X[start:end], y_: Y[start:end]})
        if i % 100 == 0:
            # 每隔一段时间计算在所有数据上的交叉熵并输出，交叉熵越小说明预测的结果和真实的结果差距越小
            total_cross_entropy = sess.run(cross_entropy, feed_dict={x: X, y_: Y})
            print("After %d training step,cross entropy on all data is %g" % (i, total_cross_entropy))
            x11[0] = x11[1]
            y11[0] = y11[1]
            x11[1] = i
            y11[1] = total_cross_entropy
            plt.plot(x11, y11)
            plt.xlabel("Step")
            plt.ylabel("Cross Entropy")
            plt.pause(0.05)
        # print(sess.run(w1))
        # print(sess.run(w2))
plt.pause(10)