题解 | 使用梯度下降的线性回归

from sklearn.linear_model import SGDRegressor
import numpy as np
import ast

def linear_regression_gradient_descent(X, y, alpha, iterations):
    # 如果 X 已经包含截距列，设置 fit_intercept=False
    m, n = X.shape
    theta = np.zeros(n)  # 初始化系数为0
    
    for _ in range(iterations):
        gradient = (1/m) * X.T @ (X @ theta - y)
        theta -= alpha * gradient
    return theta.round(4)

# 主程序
if __name__ == "__main__":
    matrix_inputx = input()
    array_y = input()
    alpha = input()
    iterations = input()

    matrix = np.array(ast.literal_eval(matrix_inputx))
    y = np.array(ast.literal_eval(array_y)).reshape(-1)  # 一维
    alpha = float(alpha)
    iterations = int(iterations)

    output = linear_regression_gradient_descent(matrix, y, alpha, iterations)
    print(output)