《机器学习实战》线性回归

　　比较简单，就是进行残差平方和最小来进行，其中这里的差为全局最优，相当于全局梯度下降法。

　　也可以参考之前写的博文：http://www.cnblogs.com/wjy-lulu/p/7759515.html

 1 import numpy as np
 2 
 3 def loadData(filename):
 4     numFeat = len(open(filename).readline().split('\t')) - 1
 5     dataMat = []
 6     labelMat = []
 7     fr = open(filename)
 8 
 9     for line in fr.readlines():
10         linArr = []
11         curline = line.strip().split('\t')
12         for i in range(numFeat):
13             linArr.append(float(curline[i]))
14         dataMat.append(linArr)
15         labelMat.append(float(curline[-1]))
16     return dataMat, labelMat
17 def standRegres(xArr,yArr):
18     xMat = np.mat(xArr)
19     yMat = np.mat(yArr).T
20     xTx = xMat.T * xMat
21     if (np.linalg.det(xTx)==0.0):#det求解的值|A|
22         print("This matrix is singular: cannot do inverse")
23         return
24     ws = xTx.I * (xMat.T*yMat)#mat.I求解矩阵的逆
25     return ws 
26 
27

main.py

 1 import logist
 2 import numpy as np
 3 import matplotlib.pyplot as plt
 4 
 5 if __name__ == '__main__':
 6     xArr, yArr = logist.loadData("ex0.txt")
 7     ws = logist.standRegres(xArr,yArr)
 8     xMat = np.mat(xArr);yMat = np.mat(yArr)
 9     xCopy = xMat.copy()
10     xCopy = np.sort(xCopy,0)#对X进行排序之后再进行画图,列排序
11     yHat = xCopy*ws
12     fg = plt.figure()
13     ax = fg.add_subplot(111)
14     ax.scatter(xMat[:,1].flatten().tolist()[0],yMat.T[:,0].flatten().tolist()[0])
15     ax.plot(xCopy[:,1],yHat)
16     plt.show()
17 
18 
19 
20     print("hello deepin")
21     print(ws);

《机器学习实战》线性回归

二.局部加权线性回归

　　理论部分请看之前笔记：http://www.cnblogs.com/wjy-lulu/p/7759515.html