Numpy/Pandas均值处理数据缺失值

系统管理员 2023-06-27 08:52 29阅读 0赞

# -*- coding: utf-8 -*-
    #-----------------------------------------------------------------------------------------------------------------------
    __Author__ = 'assasin'
    __DateTime__ = '2020/1/5 15:13'
    #-----------------------------------------------------------------------------------------------------------------------
    
    '''
    处理数据缺失值
    Numpy均值处理数据缺失值
    Pandas加载数据,重构缺失数据矩阵
    Pandas 实现均质填充
    Pandas 处理缺失值: 标量法,丢失法,忽略法,前后法
    '''
    
    import numpy as np
    import pandas as pd
    from numpy import *
    
    def loadDataSet(filepath,delim='\t'):
        fr = open(filepath)
        stringArr = [line.strip().split(delim) for line in fr.readlines()]
        #print(stringArr)
        dataArr = [list(map(float,line)) for line in stringArr]
        return mat(dataArr)
    
    
    def replaceNanwithMean(dataArr):
        numfeat = shape(dataArr)
        for i in range(numfeat[1]-1):
            meanVal = mean(dataArr[nonzero((~isnan(dataArr[:,i].A))[0],i)])
            dataArr[nonzero(isnan(dataArr[:,i].A))[0],i] = meanVal
    
        return dataArr
    
    
    
    
    
    if __name__ == '__main__':
        # 加载数据集
        dataArr = loadDataSet(r'../xxx.txt','    ')
    
        # 均值填充缺失值
        replaceNanwithMean(dataArr)
    
        datamat = loadDataSet(r'../xxx.txt','    ')
        df = pd.DataFrame(datamat)
        # 重构矩阵
        df = df.reindex(range(datamat.shape[0] + 5 ))
        # NAN 视为0
        loassVs = [df[col].mean()  for col in range(datamat.shape[1])]
        lists = [list(df[i].fillna(loassVs[i]))  for i in range(len(loassVs))]
        print(mat(lists).T)