Question

我制作了一个使用numpy进行线性回归的python脚本。当我运行它们时，出现以下错误：

警告（来自警告模块）：文件“ C：\ Users \ Moham \ Desktop \ Simple linear equation.py”，第27行 gradient_sum_squares + =梯度** 2 RuntimeWarning：在double_scalars中遇到溢出

对于出现问题的任何帮助，我们将不胜感激。

代码如下：

import numpy as np
import pandas as pd
from math import sqrt

df = pd.read_csv('Boston.csv')

def simple_linear_regression(x_points, y_points, step_size=0.01, tolerance=0.001, max_iter=1000):



 converged = False
    coefficients = np.random.rand(2)
    y = np.array(y_points)
    X = np.ones((len(x_points), 2))
    X[:, 1] = x_points
    iteration = 0

    while not converged:

        if iteration >= max_iter:
            return coefficients

        predictions = np.dot(X, coefficients)
        residual = y - predictions
        gradient_sum_squares = 0

        for i in range(len(coefficients)):
            gradient = np.dot(2, np.dot(X[:, i], residual))
            gradient_sum_squares += gradient ** 2
            coefficients[i] = coefficients[i] - np.dot(step_size, gradient)

        if sqrt(gradient_sum_squares) < tolerance:
            converged = True

        iteration += 1

    return coefficients

print(simple_linear_regression(df["TV"], df["sales"]))

RuntimeWarning：在double_scalars numpy中遇到溢出

0 个答案: