机器学习PCA实现

代码链接：

from numpy import *


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


def pca(dataMat, topNfeat=9999999):
    meanVals = mean(dataMat, axis=0)  # 求列的平均值
    meanRemoved = dataMat - meanVals  # 去平均值
    covMat = cov(meanRemoved, rowvar=0)  # 协方差矩阵
    eigVals, eigVects = linalg.eig(mat(covMat))  # 计算特征值和特征向量
    eigValInd = argsort(eigVals)  # 升序排序
    # seq[start:end:step]
    eigValInd = eigValInd[: - (topNfeat + 1):-1]  # 对N个值升序排序1
    redEigVects = eigVects[:, eigValInd]  # 对N个值升序排序2
    lowDDataMat = meanRemoved * redEigVects  # 将数据转换到新空间
    reconMat = (lowDDataMat * redEigVects.T) + meanVals
    return lowDDataMat, reconMat


def replaceNanWithMean():
    datMat = loadDataSet('secom.data', ' ')
    numFeat = shape(datMat)[1]
    for i in range(numFeat):
        # datMat[con1, col]:con1为条件.
        # 这里datMat[:,i].A是取出第i列的元素,
        # 通过~isnan()取得数字元素的引用,
        # nonzero()构造非零矩阵,所有数值元素集中在第一行,
        # datMat[nonzero(~isnan(datMat[:, i].A))[0], i]
        # 代表取得第i列中对应非零矩阵第一行中的元素
        meanVal = mean(datMat[nonzero(~isnan(datMat[:, i].A))[0], i])  # 计算平均值
        datMat[nonzero(isnan(datMat[:, i].A))[0], i] = meanVal  # 对NAN元素赋予平均值
    return datMat