利用python numpy库手动实现机器学习中的逻辑回归算法

# 手动实现逻辑回归
import numpy as np

# Sigmoid 函数
def sigmoid(z):
    return 1 / (1 + np.exp(-z))

def compute_cost(y, y_pred):
    """计算逻辑回归的成本函数"""
    m = len(y)
    cost = - (1 / m) * np.sum(y * np.log(y_pred) + (1 - y) * np.log(1 - y_pred))
    return cost

def train_logistic_regression(X, y, learning_rate=0.01, num_iterations=1000):
    # X.shape：获取特征矩阵X的形状，m是样本数量，n是特征数量。
    # weights：初始化为全零的权重向量。
    # bias：初始化为零的偏置项。
    m, n = X.shape
    weights = np.zeros(n)  # 初始化权重
    bias = 0  # 初始化偏置

	for i in range(num_iterations): # 进行num_iterations次迭代。
        # 计算预测值
        z = np.dot(X, weights) + bias
        y_pred = sigmoid(z)

        # 计算损失
        loss = compute_cost(y, y_pred)

        # 计算梯度
        dz = y_pred - y
        dw = np.dot(X.T, dz) / m
        db = np.sum(dz) / m

        # 更新权重和偏置
        weights -= learning_rate * dw # 使用学习率更新权重
        bias -= learning_rate * db # 使用学习率更新偏置。

        if i % 100 == 0:  # 每 100 次迭代打印损失
            print(f"Iteration {i}: Loss = {loss}")

    return weights, bias

# 预测函数
def predict(X, weights, bias):
    z = np.dot(X, weights) + bias
    y_pred = sigmoid(z)
    return np.round(y_pred)