ML（机器学习）感知机的权重，偏置，阈值(与门，非与门，或门，异或门)实现。

今天，学机器学习的基础，感觉还不错，有点上头。接着跟大家分享一波。

import numpy as np
import matplotlib.pyplot as plt
import time
import os

"""
#encoding="utf-8"
#Author:Mr.Pan_学狂
#finish_time:2022/9 1:03
"""

"""
感知机
**线性可分**
当对应变量的权重乘以输入的变量再求和的求和值
与阈值θ比较大小，若是小于等于阈值θ，那么就返回0
大于阈值θ，就返回1.

表达式如下：
y = ｛
    weight1 * x1 + weight2 * x2 <= θ return 0
    weight1 * x1 + weight2 * x2 > θ return 1
｝

"""

def AND(num1,num2):#与门的数据输入(0或1)
    print('这是一个与门程序1')
    weight1,weight2,theta = 0.5,0.5,0.6#weight1，weight2都是权重
    result = weight1 * num1 + weight2 * num2
    if result > theta:
        return 1
    else:
        return  0

"""
引入偏置，将上述表达式修改成
b = -θ
y = {
    weight1 * x1 + weight2 * x2 + b <= 0 return 0
    weight1 * x1 + weight2 * x2 + b > 0 return 1
}
其实，简单理解就是将θ值从不等式右侧移到左侧，
然后将-θ赋值给变量b就变成上式。
"""

def AND2(num1,num2):
    print('这是一个与门程序2')
    weight1, weight2, theta = 0.5, 0.5, 0.6  # weight1，weight2都是权重
    b = -theta#偏置
    result = weight1 * num1 + weight2 * num2 + b
    if result > 0:
        return 1
    else:
        return 0


def NAND(num1,num2):
    print('这是一个非与门程序')
    weight1,weight2,theta = 0.5,0.5,0.7
    b = -theta#偏置
    result = weight1 * num1 + weight2 * num2 + b
    if result > 0:
        return 0
    else:
        return 1

def NAND2(num1,num2):
    print('这是一个非与门程序2')
    theta = 0.7#阈值
    weights = np.array([0.5,0.5])#因为是一维数组(向量)
    InX = np.array([num1,num2])#将输入的变量构建成一个向量
    result = weights * InX#运用向量的乘法
    # print(result)
    result2 = np.sum(result)#对结果求和
    b = -theta  # 偏置量
    result3 = result2 + b#再加上偏置量
    if result3 > 0:
        return 0
    else:
        return 1

def OR(inx1,inx2):
    print('这是一个或门程序')#跟与门程序同理，不过偏置量设置不同。
    weights = np.array([0.5,0.5])
    theta = 0.3#阈值θ
    b = -theta#偏置
    InX = np.array([inx1,inx2])
    result = InX * weights
    result2 = np.sum(result)
    result3 = result2 + b
    if result3 > 0:
        return 1
    else:
        return 0

"""
异或门是属于线性不可分
所以，简单的感知机不能正确划分。
"""

def XOR(inX1,inX2):
    """
    通过构建2层感知机，
    将非与门和与门合并，
    构建出异或门
    """
    # pass
    print('这是一个异或门程序')
    weights = np.array([0.5,0.5])
    InX = np.array([inX1,inX2])
    theta = 0.6#阈值
    b = -theta#偏置
    result = weights * InX
    result2 = np.sum(result)
    if result2 == 0:
        return 0
    else:
        result3 = result2 + b
        if result3 < 0:
            return 1
        else:
            return 0
            
def XOR2(inx1,inx2):
    print('这是一个异或门程序2')
    num1 = NAND(inx1,inx2)
    num2 = OR(inx1,inx2)
    y = AND(num1,num2)
    return y


if __name__ == '__main__':
    #pass
    for num in range(2):
        print(AND(num,num))
    print(AND(0,1))
    print(AND(1,0))
    print()
    print()
    for num in range(2):
        print(AND2(num, num))
    print(AND2(0, 1))
    print(AND2(1, 0))
    print()
    print()
    for num in range(2):
        print(NAND(num, num))
    print(NAND(0, 1))
    print(NAND(1, 0))
    print()
    print()
    for num in range(2):
        print(NAND2(num, num))
    print(NAND2(0, 1))
    print(NAND2(1, 0))
    print()
    print()
    for num in range(2):
        print(OR(num, num))
    print(OR(0, 1))
    print(OR(1, 0))
    print()
    print()
    for n in range(2):
        print(XOR(n,n))
    print(XOR(1,0))
    print(XOR(0,1))

运行结果：



最后，感谢大家前来观看鄙人的文章，文中或有诸多不妥之处，还望指出和海涵。