机器学习：梯度下降法与线性回归

根据麦子学院彭亮老师的机器学习梯度下降法的python程序而写

#coding=utf-8
import numpy as np
import random


# 梯度下降算法
def gradientDescent(x, y, theta, alpha, m, numIteration):  # 输入实例、分类标签、要学习的参数、学习率、实例个数、迭代次数
    xTrans = x.transpose()  # 矩阵的转置
    for i in range(0, numIteration):
        hypothsis = np.dot(x, theta)
        # print 'hypothsis',hypothsis
        loss = hypothsis - y
        cost = np.sum(loss ** 2) / (2 * m)
        print("Iteration %d / Cost:%f" % (i, cost))
        gradient = np.dot(xTrans, loss) / m
        theta = theta - alpha * gradient  # 更新法则
    return theta


# 创建数据，用作测试
def genData(numPoints, bais, variance):
    '''

    :param numPoints: 样本个数
    :param bais: 偏差
    :param variance: 方差
    :return:
    '''
    x = np.zeros(shape=(numPoints, 2))  # 全零矩阵，行数为样本个数，列数为2
    y = np.zeros(shape=numPoints)  # 标签，零向量。行数为样本个数，列数为1
    for i in range(0, numPoints):
        x[i][0] = 1  # 所有行第1列为：1
        x[i][1] = i  # 所有行第2列为：行的数目
        y[i] = (i + bais) + random.uniform(0, 1) * variance
    return x, y


x, y = genData(100, 25, 10)
print("x:", x)
print("y:", y)
m, n = np.shape(x)
n_y = np.shape(y)
print "m: %s n: %s "  %(str(m),str(n))
print "n_y: %s"   %(str(n_y))
numIteration = 120000
alpha = 0.0005
theta = np.ones(n)  # 初始化theta  1*n
theta = gradientDescent(x, y, theta, alpha, m, numIteration)
print theta

结果：

[ 30.56233795   0.99501898]