sklearn-LR_prov.py 文件源码

def loadDataSet(filename):
    df = pd.read_excel(filename,sheetname=[1], header=None, skiprows=1)[1]
    df = df.fillna(0)
    df[2] = df[2]/100000000  # ?????
#    zeros = df[df[0]==0]
#    df = df.drop(zeros.index,axis=0)
    df[2] = standard(df[2]) #?????

    df[3] = standard(df[3]) #????
    df[4] = standard(df[4]) #????
    df[6] = standard(df[6]) #??????
    df[9] = df[9].apply(map_01)  # ?????
    df[11] = df[11].apply(map_01) # ??????
    df[14] = standard(df[14]) #???????
    df[15] = standard(df[15].apply(map_rate)) #???????
    df[16] = standard(df[16].apply(map_sub_rate)) #?????????
    df[17] = standard(df[17]) #??????
    prov_coding,province_dict = transcoding(df[10]) # province

    enter_coding,enter_dict = transcoding(df[13]) # enterprise
    target = df[2]
    data = df[[3,4,6,9,11,14,15,16,17]]
    data = pd.concat([data,prov_coding],axis=1)
    data = pd.concat([data,enter_coding],axis=1)

    import seaborn as sns
#    sns.pairplot(df, x_vars=[3,4,6,9,11,14,15,16,17], y_vars=2, size=7, aspect=0.8, kind='reg')
    return mat(data),mat(target).T