pytorch（一）简单回归笔记

import numpy as np

# y = wx + b ===》 模型E(w,b) = Σ(yi-wxi-b)²
# 计算总的平均error
def compute_error_for_line_given_points(b, w, points):
    totalError = 0
    # 循环100次
    for i in range(0, len(points)):
        # 每次取横坐标
        x = points[i, 0]
        # 每次取纵坐标
        y = points[i, 1]
        # loss= Σ(Wx+b - y)²
        totalError += (y - (w * x + b)) ** 2
    #     对总的error求一个平均，返回总的平均error
    return totalError / float(len(points))

# 定义一个计算梯度的函数：入口参数：b当前值，W当前值，点集合，学习率
def step_gradient(b_current, w_current, points, learningRate):
    b_gradient = 0
    w_gradient = 0
    N = float(len(points))
    for i in range(0, len(points)):
        x = points[i, 0]
        y = points[i, 1]
        # 用最小二乘法求参数估计b最小 对b求偏导    ∂E(w,b)/∂b= -2Σ(y-(wx+b))
        b_gradient += -(2/N) * (y - ((w_current * x) + b_current))
        # 对w求偏导 ∂E(w,b)/∂w = -2Σ(y-(wx+b))x
        # 所有梯度累加时除以N，以便结果不用再做平均
        w_gradient += -(2/N) * x * (y - ((w_current * x) + b_current))
    new_b = b_current - (learningRate * b_gradient)
    new_m = w_current - (learningRate * w_gradient)
    return [new_b, new_m]

# 循环迭代W，b，入口参数：（x，y）点集合，初始b,初始W，学习率，迭代次数：
def gradient_descent_runner(points, starting_b, starting_m, learning_rate, num_iterations):
    b = starting_b
    m = starting_m
    for i in range(num_iterations):
        b, m = step_gradient(b, m, np.array(points), learning_rate)
    return [b, m]

def run():
    # genfromtxt主要执行两个循环运算。
    # 第一个循环将文件的每一行转换成字符串序列。
    # 第二个循环将每个字符串序列转换为相应的数据类型。
    # 将数据读入，以‘，’分割，处理为数组类型
    points = np.genfromtxt("data.csv", delimiter=",")
    # print(points)
    # 学习率0.0001
    learning_rate = 0.0001
    initial_b = 0 # initial y-intercept guess  b
    initial_m = 0 # initial slope guess  w
    num_iterations = 1000
    print("Starting gradient descent at b = {0}, m = {1}, error = {2}"
          .format(initial_b, initial_m,
                  compute_error_for_line_given_points(initial_b, initial_m, points))
          )
    print("Running...")
    [b, m] = gradient_descent_runner(points, initial_b, initial_m, learning_rate, num_iterations)
    print("After {0} iterations b = {1}, m = {2}, error = {3}".
          format(num_iterations, b, m,
                 compute_error_for_line_given_points(b, m, points))
          )

if __name__ == '__main__':
    run()