简单逻辑回归的Python Vanilla代码

import numpy as np

def sigmoid(z):
    """
    Sigmoid函数的实现
    """
    return 1 / (1 + np.exp(-z))

def logistic_regression(X, y, num_iterations, learning_rate):
    """
    简单逻辑回归模型的训练函数
    参数：
        X: 特征矩阵，shape为 (样本数, 特征数)
        y: 标签向量，shape为 (样本数, )
        num_iterations: 迭代次数
        learning_rate: 学习率
    返回：
        theta: 训练得到的模型参数，shape为 (特征数, )
    """
    m, n = X.shape
    theta = np.zeros(n)  # 初始化模型参数

    for i in range(num_iterations):
        z = np.dot(X, theta)  # 计算线性组合
        h = sigmoid(z)  # 通过sigmoid函数得到预测概率
        gradient = np.dot(X.T, (h - y)) / m  # 计算梯度
        theta -= learning_rate * gradient  # 更新模型参数

    return theta

# 使用示例数据进行训练和预测
X_train = np.array([[1, 2], [2, 3], [3, 4], [4, 5]])
y_train = np.array([0, 0, 1, 1])
theta = logistic_regression(X_train, y_train, num_iterations=1000, learning_rate=0.1)
print("模型参数：", theta)

# 预测新样本
X_test = np.array([[5, 6], [6, 7]])
z = np.dot(X_test, theta)
h = sigmoid(z)
predictions = (h >= 0.5).astype(int)
print("预测结果：", predictions)