随机梯度下降法底层

import numpy as np
import matplotlib.pyplot as plt

数据集创建

m= 100000

x=np.random.normal(size=m)
X=x.reshape(-1,1)
y=4.*x+3.+np.random.normal(0,3,size=m)

其实效用不大，但不定义，又违背了梯度下降算法的要求
求导
求y值

def J(theta,X_b,y):
    try:
        return np.sum((y - X_b.dot(theta))**2 )/len(X_b)
    except:
        return float('inf')
    
def dj_sgd(theta,X_b_i,y_i):
    return X_b_i.T.dot(X_b_i.dot(theta)-y_i )*2.

底层实现

def sgd(X_b,y,initial_theta,n_iters=5):
    '''
    initial_theta 样本特征比例
    
    n_iters   对于整个样本遍历的轮数
    '''
    t0=5
    t1=50
    m=len(X_b)
    
    def learning_rate(t):
        return t0 /(t+t1)

    theta = initial_theta
    for cur_iter in range(n_iters):
        indexs = np.random.permutation(m)   #随机索引乱序  达到能够全部都能够遍历
        X_b_new=X_b[indexs]  #用新的索引来进行排序
        y_b_new=y[indexs]
        for i in range(m):
            gradient = dj_sgd(theta,X_b[i],y[i])
            theta = theta - learning_rate(cur_iter*m+i) * gradient
    return theta

x_b=np.hstack([np.ones((len(X),1)),X])
initial_theta=np.zeros(x_b.shape[1])
theta=sgd(x_b,y,initial_theta)
theta