Python实现PCA

糖逗

参考书籍：《机器学习实战》

import numpy as np

def loadDataSet(fileName, delim = '\t'):
    fr = open(fileName)
    stringArr = [line.strip().split(delim) for line in fr.readlines()]
    datArr = [list(map(float,line)) for line in stringArr]
    return np.mat(datArr)

def pca(dataMat, topNfeat = 9999999):
    meanVals = np.mean(dataMat, axis = 0)
    meanRemoved = dataMat - meanVals 
    covMat = np.cov(meanRemoved, rowvar = 0)#rowvar=0表示传入的一行表示一个样本
    eigVals, eigVects = np.linalg.eig(np.mat(covMat))
    eigValInd = np.argsort(eigVals)           
    eigValInd = eigValInd[:-(topNfeat+1):-1]  
    redEigVects = eigVects[:,eigValInd]       
    lowDDataMat = meanRemoved * redEigVects#所有点在新的一组基上的值，无方向
    reconMat = (lowDDataMat * redEigVects.T) + meanVals#所有点在新的一组基上的投影，有方向的
    return lowDDataMat, reconMat


if __name__ == '__main__':
    dataMat = loadDataSet(r'C:\...\testSet.txt')
    lowData, reconMat = pca(dataMat, 1)
    
    import matplotlib.pyplot as plt
    fig = plt.figure()
    ax = fig.add_subplot(111)
    ax.scatter(dataMat[:, 0].flatten().A[0], dataMat[:, 1].flatten().A[0],
               marker = '^', s = 90)
    ax.scatter(reconMat[:, 0].flatten().A[0], reconMat[:, 1].flatten().A[0],
               marker = 'o', s = 50, c = 'red')
    

糖逗