52、逻辑回归_1 for i in range(len(y1)) if y[i] == y1[i] 1 else

import random as rd
import matplotlib.pyplot as plt
import functools
import math
import numpy as np
 
def sigmoid(x):
    if x >= 500:
        return 0
    if x <= -500:
        return 1
    return 1/(1+math.exp(-x))
 
def sample_generate(n, low, sup, w, e):
    X = [[(sup-low) * rd.random() + low if j < len(w)-1 else 1 for j in range(len(w))] for i in range(n)]
    Y = [round(sigmoid(sum([w[j]*X[i][j] for j in range(len(w))])+e*(rd.random()-0.5))) for i in range(n)]
    return X,Y
 
def LR_2_plot(X, Y):
 
    X = [X[i][0] for i in range(len(X))]
    c = ["r" if Y[i] == 1 else "b" for i in range(len(Y))]
    #点、线展示
    plt.scatter(X, Y, edgecolors=c)
    plt.show()
 
def f(w, X, Y):
    return sum([math.pow((sigmoid(sum([X[i][j]*w[j] for j in range(len(w))])) - Y[i]),2) for i in range(len(X))])
 
def mod(dw):
    return math.sqrt(sum([dw[i]*dw[i] for i in range(len(dw))]))
 
# r为放缩尺度, step_min为最小步长
def Gradient(X, Y, r = 0.8, step_min = 0.00001):
    n = len(X)
    d = len(X[0])
    w = [1 for i in range(d)]
 
    while True:
        print(w)
        left = 0
        right = 1
        dw = [sum([X[j][i]*(sigmoid(sum([X[j][k]*w[k] for k in range(d)])) - Y[j])*(sigmoid(sum([X[j][k]*w[k] for k in range(d)])) - 1) for j in range(n)]) for i in range(d)]
        # 先采用放缩法确定寻优区间
        while f([w[i]+right*dw[i] for i in range(d)], X, Y) < f(w, X, Y):
            right = right*(2-r)
        # 采用三分法确定步长
        mid1 = left*2/3 + right/3
        mid2 = left/3 + right*2/3
        while abs(left - right)*mod(dw) > step_min:
            if f([w[i]+mid1*dw[i] for i in range(d)], X, Y) < f([w[i]+mid2*dw[i] for i in range(d)], X, Y):
                right = mid2
            else:
                left = mid1
            mid1 = left * 2 / 3 + right / 3
            mid2 = left / 3 + right * 2 / 3
 
        if left*mod(dw) < step_min:
            break
        w = [w[i] + left * dw[i] for i in range(d)]
 
    return w
 
 
if __name__ == "__main__":
 
 
    X, Y = sample_generate(n = 1000, low = -10, sup = 10, w = [1,2,3,4,5], e = 5)
    # X = [[8.413821508016959, 1], [8.899549892638564, 1], [-2.3498826552587904, 1], [9.602673657236384, 1], [9.195780070565458, 1], [-5.545278156033005, 1], [-6.34749638138797, 1], [5.392609937972598, 1], [-8.072733305295914, 1], [-7.228355148125498, 1]]
    # Y = [1, 1, 0, 1, 1, 0, 0, 1, 0, 0]
    print(X)
    print(Y)
 
    #样本画图，仅限w为2维
    # LR_2_plot(X, Y)
 
    # 梯度下降法
    w = Gradient(X, Y)
    print("\nw: ", w)
 
    Y1 = [round(sigmoid(sum([w[j]*X[i][j] for j in range(len(w))]))) for i in range(len(X))]
    print("回归准确率:", sum([1 if Y[i] == Y1[i] else 0 for i in range(len(Y))])/len(Y))