简单的人工神经网络的实现

import numpy as np
import os
import time

def sigmoid(x):                                             # 定义激活函数
    return 1 / (1 + np.exp(-x))

def sigmoid_derivative(x):                                  # 定义激活函数的导数
    return x * (1 - x)

inputs = np.array(
   [
      [0, 0, 1],
      [1, 1, 0],
      [0, 0, 0],
      [1, 1, 1],                                            #样本
      [1, 0, 1],
      [0, 1, 1]
   ])                                                       #输入数据

outputs = np.array([
   [0],
   [1],
   [0],
   [1],                                                     #样本标注
   [1],
   [0]
   ])                                                       # 目标输出

np.random.seed(1)                                           # 随机初始化权重
weights = 2 * np.random.random((inputs.shape[1], 1)) - 1    # inputs的列数
i = 0

for _ in range(10000000):                                   # 训练神经网络
    input_layer = inputs

    weighted_sum = np.dot(input_layer, weights)             # 将输入向量与权重矩阵相乘
    output_layer = sigmoid(weighted_sum)                    # 应用激活函数（sigmoid）

    test = np.array([
      [0, 1, 0], [0, 1, 1], [0, 0, 0], [0, 0, 1],
      [1, 1, 0], [1, 1, 1], [1, 0, 0], [1, 0, 1]])
    output = sigmoid(np.dot(test, weights))# 求缺失的

    error = outputs - output_layer                          # 计算误差
   
    adjustments = error * sigmoid_derivative(output_layer)  # 更新权重
    weights += np.dot(input_layer.T, adjustments)

    i = i + 1  
    if not i % 1:    
      print(
      "样本标注:", "\n", outputs, "\n", "\n",
      "样本:", "\n", input_layer, "\n", "\n",
      "权重和:", "\n", weighted_sum, "\n", "\n",      
      "输出层:", "\n", output_layer, "\n", "\n",
      "误差:", "\n", error, "\n", "\n",
      "调整:", "\n", adjustments, "\n", "\n",
      "权重:", "\n", weights, "\n", "\n", "\n",
      "重复次数:", i, "\n",
      "输出:", "\n", output, "\n"
      )

      test=np.array([[i/1000.0, i/1000.0, i/1000.0, i/1000.0]])
      print("test:\n", sigmoid(test))
 
      time.sleep(0.1)                                       # 等待1秒
      os.system('cls' if os.name == 'nt' else 'clear')       # 根据操作系统选择清屏命令    
    
# 测试神经网络
print("输出结果：", output_layer)

好像能判断一个向量第一维度是不是1