【吴恩达机器学习】ex2 : logistic regression ——Python

最新推荐文章于 2025-08-11 22:22:04 发布

鱼儿冒个泡

最新推荐文章于 2025-08-11 22:22:04 发布

阅读量168

点赞数

CC 4.0 BY-SA版权

分类专栏：吴恩达机器学习作业文章标签： python 机器学习

本文链接：https://blog.csdn.net/hhhenjoy/article/details/119983733

吴恩达机器学习作业专栏收录该内容

4 篇文章

订阅专栏

这段Python代码展示了如何利用梯度下降法实现逻辑回归。首先定义了sigmoid激活函数，然后实现梯度下降算法进行参数更新。数据集从'ex2data1.txt'文件加载，通过散点图展示数据分布，并绘制拟合曲线。最后，给出训练得到的模型参数。

摘要生成于 C知道，由 DeepSeek-R1 满血版支持，前往体验 >

logistic regression

Python代码：

import numpy as np
import pandas as pd
import matplotlib.pyplot as plt

def sigmond(x):
    sig = 1 / (1 + np.exp(-x))
    return sig

def gradient_descend(X, Y, init_theta, alpha, iter_nums):
    theta = init_theta
    for i in range(0, iter_nums):
        hyp = sigmond(np.dot(np.transpose(theta), X)) - np.transpose(Y)
        hyp = np.dot(hyp, np.transpose(X))
        theta -= hyp * alpha / m
    return theta

# load the data
filepath = r'C:\Users\LENOVO\OneDrive - 东南大学\桌面\机器学习\吴恩达作业\python\ex2\data\ex2data1.txt'
dataset = np.loadtxt(filepath, delimiter=',', usecols=(0, 1, 2))

[m, n] = np.shape(dataset)
x1 = dataset[:, 0]
x2 = dataset[:, 1]
y = dataset[:, 2]
X = np.c_[np.ones(m), x1, x2]
X = np.transpose(X)
Y = y

# scatter
pos = (y == 0)
neg = (y == 1)
plt.plot(x1[pos], x2[pos], 'r+')
plt.plot(x1[neg], x2[neg], 'go')

init_theta = np.float64(np.transpose(np.array([0, 0, 0])))
alpha = 0.001
iter_nums = 1000000

theta = gradient_descend(X, Y, init_theta, alpha, iter_nums)
print(theta)

# fitted curve
fitted_x2 = np.ones((m, 1), dtype=float, order='C') * theta[1]
for i in range(0, m):
    fitted_x2[i, 0] = (fitted_x2[i, 0] * x1[i] + theta[0]) / theta[2] * (-1)
plt.plot(x1, fitted_x2, 'b-')
plt.show()