Skip to content
/ AugustBetty1.0.2 Public

评分卡 逻辑回归 最优分组 金融风控 信用等级 屁屁和铭仔的数据之路 AugustBetty

11 stars 3 forks Branches Tags Activity
Star
Notifications You must be signed in to change notification settings

XueRenJing/AugustBetty1.0.2

BranchesTags

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

6 Commits
.idea
.idea
 
 
AugustBetty
AugustBetty
 
 
README.md
README.md
 
 
README.txt
README.txt
 
 
setup.py
setup.py
 
 
深度学习.jpg
深度学习.jpg
 
 

Repository files navigation

AugustBetty能做什么

这个包是由微信公众号"屁屁和铭仔的数据之路"博主铭仔和屁屁一起开发的，目的是为了将逻辑回归模型开发流程规范化，

安装

pip install AugustBetty

导入

from AugustBetty.AugustBetty import ModelMain

一句话建模

 data = pd.read_csv(r"F:\论文-风控\train.csv")  #读取训练集
 m=MakeModel(data,"id","isDefault",0.02,5,512254,drop_value=["issueDate","earliesCreditLine"])
 m.model_main()

v1.0.0的功能如下：

数值变量的最优分组

输入指定的参数即可进行最优分组 分组结果包含以下四项： 1. 已经包含最优分组后的变量的数据集（在原始数据集上增加分组后的变量） 2. 最优分组后的变量名以及没有进行最优分组的变量（进行是组数较少，不进行最优分组） 3. 参与最优分组的变量的映射口径以及转换之后的变量名 4. 计算输入变量中转换了的变量的iv以及woe.

字符变量的最优分组

   1. 已经包含最优分组后的变量的数据集（在原始数据集上增加分组后的变量）
   2. 最优分组后的变量名以及没有进行最优分组的变量（进行是组数较少，不进行最优分组）
   3. 参与最优分组的变量的映射口径以及转换之后的变量名
   4. 计算输入变量中转换了的变量的iv以及woe.

逻辑回归模型的开发

  1. 训练集测试集的自动分区
  2. 自动识别数据集中的字符以及数值变量进行最优分组
  3. 自动转woe
  4. 自动根据iv设定阈值，筛选出高于iv值的变量进入模型中拟合
  5. 测试集根据分组映射woe,计算测试集的ks&auc,模型最终结果是模型的summary信息，以及测试集和训练集的ks&auc结果。

V1.1.0的功能规划如下：

拆分出细致的单独功能： 1、单独的函数转原始数据 2、增加验证集（在原始数据集上使用时间参数，划分验证集，跑模型得出评估指标） 3、在输入的路径下，输出模型文件

V1.2.0的功能规划如下：

开发自动部署模型的函数，输入模型位置，以及数据转换文件，输出模型api,以支持部分公司微服务的需求

如有报错，或者其他需求，请发邮件到：watchmans@qq.com 希望能不断的完善这个包，这个包现阶段功能相对单一。 欢迎大家关注公众号：屁屁和铭仔的数据之路

数值变量分组调用

执行命令：gg=num_group_chi_main(train_data,id_name,y,initial_group_num,group_num)

train_data:原始数据集

id_name:主键名字，例如"idno"

y:y变量名,例如"isbad"

initial_group_num:粗分组组数，这里的逻辑是将数值变量先粗分组再细分组的方式进行最优分组，一般设定为20-50，看数据颗粒度

group_num：最优分组组数，一般设定为4-6组

 dict_keys(['data', 'var_dict_total', 'var_bin_list', 'iv_dict'])

data:增加新变量之后的数据集。原始变量没有改动。只对数值变量做分组。

{'data':             id  loanAmnt  term  ...  n13_group_Bin  n14_group n14_group_Bin
531009  531009   18000.0     5  ...          Bin 0          1         Bin 0
370965  370965    3000.0     5  ...          Bin 0          1         Bin 0
212592  212592   23600.0     3  ...          Bin 0          1         Bin 0
610540  610540   20000.0     3  ...          Bin 0          1         Bin 0
400590  400590   24250.0     3  ...          Bin 0          1         Bin 0
      ...       ...   ...  ...            ...        ...           ...
603509  603509    5600.0     3  ...          Bin 0          7         Bin 4
157628  157628    7200.0     3  ...          Bin 0          7         Bin 4
147487  147487   10000.0     3  ...          Bin 0          7         Bin 4
719781  719781   16000.0     3  ...          Bin 0          7         Bin 4
735444  735444   20000.0     3  ...          Bin 0          7         Bin 4

[800000 rows x 119 columns],

var_dict_total:保存的是变量的分组口径

'var_dict_total': {'loanAmnt': {'bin': {'Bin 0': [-9999.0, 8000.0],
   'Bin 1': [8000.0, 10000.0],
   'Bin 2': [10000.0, 15000.0],
   'Bin 3': [15000.0, 28000.0],
   'Bin 4': [28000.0, 10000000000]},
  'value_name': 'loanAmnt_group_Bin'},
 'interestRate': {'bin': {'Bin 0': [-9999.0, 7.97],
   'Bin 1': [7.97, 12.29],
   'Bin 2': [12.29, 15.99],
   'Bin 3': [15.99, 22.15],
   'Bin 4': [22.15, 10000000000]},
  'value_name': 'interestRate_group_Bin'},
 'installment': {'bin': {'Bin 0': [-9999.0, 248.45],
   'Bin 1': [248.45, 324.3],
   'Bin 2': [324.3, 10000000000]},
  'value_name': 'installment_group_Bin'},
 'employmentTitle': {'bin': {'Bin 0': [-9999.0, 54.0],
   'Bin 1': [54.0, 169034.0],
   'Bin 2': [169034.0, 10000000000]},
  'value_name': 'employmentTitle_group_Bin'},
 'homeOwnership': {'bin': {'Bin 0': [-9999.0, 1], 'Bin 1': [1, 10000000000]},
  'value_name': 'homeOwnership_group_Bin'},
 'annualIncome': {'bin': {'Bin 0': [-9999.0, 45600.0],
   'Bin 1': [45600.0, 60000.0],
   'Bin 2': [60000.0, 85000.0],
   'Bin 3': [85000.0, 125000.0],
   'Bin 4': [125000.0, 10000000000]},
  'value_name': 'annualIncome_group_Bin'},
 'purpose': {'bin': {'Bin 0': [-9999.0, 2], 'Bin 1': [2, 10000000000]},
  'value_name': 'purpose_group_Bin'},
 'postCode': {'bin': {'Bin 0': [-9999.0, 19.0],
   'Bin 1': [19.0, 262.0],
   'Bin 2': [262.0, 10000000000]},
  'value_name': 'postCode_group_Bin'},
 'regionCode': {'bin': {'Bin 0': [-9999.0, 13], 'Bin 1': [13, 10000000000]},
  'value_name': 'regionCode_group_Bin'},
 'dti': {'bin': {'Bin 0': [-9999.0, 14.18],
   'Bin 1': [14.18, 21.25],
   'Bin 2': [21.25, 25.69],
   'Bin 3': [25.69, 29.78],
   'Bin 4': [29.78, 10000000000]},
  'value_name': 'dti_group_Bin'},
 'delinquency_2years': {'bin': {'Bin 0': [-9999.0, 1.0],
   'Bin 1': [1.0, 2.0],
   'Bin 2': [2.0, 10000000000]},
  'value_name': 'delinquency_2years_group_Bin'},
 'ficoRangeLow': {'bin': {'Bin 0': [-9999.0, 670.0],
   'Bin 1': [670.0, 690.0],
   'Bin 2': [690.0, 720.0],
   'Bin 3': [720.0, 760.0],
   'Bin 4': [760.0, 10000000000]},
  'value_name': 'ficoRangeLow_group_Bin'},
 'ficoRangeHigh': {'bin': {'Bin 0': [-9999.0, 674.0],
   'Bin 1': [674.0, 694.0],
   'Bin 2': [694.0, 724.0],
   'Bin 3': [724.0, 764.0],
   'Bin 4': [764.0, 10000000000]},
  'value_name': 'ficoRangeHigh_group_Bin'},
 'openAcc': {'bin': {'Bin 0': [-9999.0, 7.0],
   'Bin 1': [7.0, 10.0],
   'Bin 2': [10.0, 17.0],
   'Bin 3': [17.0, 22.0],
   'Bin 4': [22.0, 10000000000]},
  'value_name': 'openAcc_group_Bin'},
 'pubRec': {'bin': {'Bin 0': [-9999.0, 1.0], 'Bin 1': [1.0, 10000000000]},
  'value_name': 'pubRec_group_Bin'},
 'pubRecBankruptcies': {'bin': {'Bin 0': [-9999.0, 0.0],
   'Bin 1': [0.0, 1.0],
   'Bin 2': [1.0, 10000000000]},
  'value_name': 'pubRecBankruptcies_group_Bin'},
 'revolBal': {'bin': {'Bin 0': [-9999.0, 22682.0],
   'Bin 1': [22682.0, 43290.0],
   'Bin 2': [43290.0, 10000000000]},
  'value_name': 'revolBal_group_Bin'},
 'revolUtil': {'bin': {'Bin 0': [-9999.0, 18.1],
   'Bin 1': [18.1, 33.4],
   'Bin 2': [33.4, 52.1],
   'Bin 3': [52.1, 74.9],
   'Bin 4': [74.9, 10000000000]},
  'value_name': 'revolUtil_group_Bin'},
 'totalAcc': {'bin': {'Bin 0': [-9999.0, 16.0],
   'Bin 1': [16.0, 22.0],
   'Bin 2': [22.0, 10000000000]},
  'value_name': 'totalAcc_group_Bin'},
 'title': {'bin': {'Bin 0': [-9999.0, 1.0],
   'Bin 1': [1.0, 10.0],
   'Bin 2': [10.0, 10000000000]},
  'value_name': 'title_group_Bin'},
 'n0': {'bin': {'Bin 0': [-9999.0, 0.0], 'Bin 1': [0.0, 10000000000]},
  'value_name': 'n0_group_Bin'},
 'n1': {'bin': {'Bin 0': [-9999.0, 2.0],
   'Bin 1': [2.0, 3.0],
   'Bin 2': [3.0, 4.0],
   'Bin 3': [4.0, 6.0],
   'Bin 4': [6.0, 10000000000]},
  'value_name': 'n1_group_Bin'},
 'n2': {'bin': {'Bin 0': [-9999.0, 2.0],
   'Bin 1': [2.0, 4.0],
   'Bin 2': [4.0, 7.0],
   'Bin 3': [7.0, 10.0],
   'Bin 4': [10.0, 10000000000]},
  'value_name': 'n2_group_Bin'},
 'n3': {'bin': {'Bin 0': [-9999.0, 2.0],
   'Bin 1': [2.0, 4.0],
   'Bin 2': [4.0, 7.0],
   'Bin 3': [7.0, 10.0],
   'Bin 4': [10.0, 10000000000]},
  'value_name': 'n3_group_Bin'},
 'n4': {'bin': {'Bin 0': [-9999.0, 0.0],
   'Bin 1': [0.0, 4.0],
   'Bin 2': [4.0, 6.0],
   'Bin 3': [6.0, 10.0],
   'Bin 4': [10.0, 10000000000]},
  'value_name': 'n4_group_Bin'},
 'n5': {'bin': {'Bin 0': [-9999.0, 4.0], 'Bin 1': [4.0, 10000000000]},
  'value_name': 'n5_group_Bin'},
 'n6': {'bin': {'Bin 0': [-9999.0, 1.0],
   'Bin 1': [1.0, 23.0],
   'Bin 2': [23.0, 10000000000]},
  'value_name': 'n6_group_Bin'},
 'n7': {'bin': {'Bin 0': [-9999.0, 4.0],
   'Bin 1': [4.0, 7.0],
   'Bin 2': [7.0, 10.0],
   'Bin 3': [10.0, 14.0],
   'Bin 4': [14.0, 10000000000]},
  'value_name': 'n7_group_Bin'},
 'n8': {'bin': {'Bin 0': [-9999.0, 5.0], 'Bin 1': [5.0, 10000000000]},
  'value_name': 'n8_group_Bin'},
 'n9': {'bin': {'Bin 0': [-9999.0, 3.0],
   'Bin 1': [3.0, 5.0],
   'Bin 2': [5.0, 7.0],
   'Bin 3': [7.0, 10.0],
   'Bin 4': [10.0, 10000000000]},
  'value_name': 'n9_group_Bin'},
 'n10': {'bin': {'Bin 0': [-9999.0, 3.0],
   'Bin 1': [3.0, 8.0],
   'Bin 2': [8.0, 10.0],
   'Bin 3': [10.0, 22.0],
   'Bin 4': [22.0, 10000000000]},
  'value_name': 'n10_group_Bin'},
 'n11': {'bin': {'Bin 0': [-9999.0, 0.0], 'Bin 1': [0.0, 10000000000]},
  'value_name': 'n11_group_Bin'},
 'n12': {'bin': {'Bin 0': [-9999.0, 0.0], 'Bin 1': [0.0, 10000000000]},
  'value_name': 'n12_group_Bin'},
 'n13': {'bin': {'Bin 0': [-9999.0, 0.0], 'Bin 1': [0.0, 10000000000]},
  'value_name': 'n13_group_Bin'},
 'n14': {'bin': {'Bin 0': [-9999.0, 0.0],
   'Bin 1': [0.0, 1.0],
   'Bin 2': [1.0, 2.0],
   'Bin 3': [2.0, 3.0],
   'Bin 4': [3.0, 10000000000]},
  'value_name': 'n14_group_Bin'}},

var_bin_list：保存可用于如模型拟合的变量（逻辑回归本质上入模的变量都应该是数值变量，这里的list保存的时候分组好的变量以及原本的分组组已经满足组数（一般设定5组）的变量list,输出这个的原因是后续好循环转换woe,无需手工把变量敲出来。）

'var_bin_list': ['loanAmnt_group_Bin',
'interestRate_group_Bin',
'installment_group_Bin',
'employmentTitle_group_Bin',
'homeOwnership_group_Bin',
'annualIncome_group_Bin',
'purpose_group_Bin',
'postCode_group_Bin',
'regionCode_group_Bin',
'dti_group_Bin',
'delinquency_2years_group_Bin',
'ficoRangeLow_group_Bin',
'ficoRangeHigh_group_Bin',
'openAcc_group_Bin',
'pubRec_group_Bin',
'pubRecBankruptcies_group_Bin',
'revolBal_group_Bin',
'revolUtil_group_Bin',
'totalAcc_group_Bin',
'title_group_Bin',
'n0_group_Bin',
'n1_group_Bin',
'n2_group_Bin',
'n3_group_Bin',
'n4_group_Bin',
'n5_group_Bin',
'n6_group_Bin',
'n7_group_Bin',
'n8_group_Bin',
'n9_group_Bin',
'n10_group_Bin',
'n11_group_Bin',
'n12_group_Bin',
'n13_group_Bin',
'n14_group_Bin',
'term',
'verificationStatus',
'initialListStatus',
'applicationType',
'policyCode'],

iv_dict：保存是的输入的数据集中可最优分组的变量，转换后的iv,woe,以及已经是足够组数的变量的iv,woe

'iv_dict': {'term': {'iv': 0.17263507884380802,
 'woe': {3: -0.2686182828015269, 5: 0.6520081414240493}},
'verificationStatus': {'iv': 0.05451891242377009,
 'woe': {0: -0.3672196958578996,
  1: 0.06087198832351855,
  2: 0.22488431630897235}},
'initialListStatus': {'iv': 0.0003419546491836722,
 'woe': {0: 0.015549994613464051, 1: -0.021991286981396775}},
'applicationType': {'iv': 0.001906255272074922,
 'woe': {0: -0.006394473785834787, 1: 0.2981570883254602}},
'policyCode': {'iv': 0.0, 'woe': {1.0: 0.0}},
'loanAmnt_group_Bin': {'iv': 0.03450643699626254,
 'woe': {'Bin 0': -0.26698353031170635,
  'Bin 1': -0.10585087483552284,
  'Bin 2': 0.05207841143835551,
  'Bin 3': 0.15704067216706727,
  'Bin 4': 0.2587985257714207}},
'interestRate_group_Bin': {'iv': 0.432844128756394,
 'woe': {'Bin 0': -1.4139767754271722,
  'Bin 1': -0.47444483191986947,
  'Bin 2': 0.12600617373518758,
  'Bin 3': 0.6329403904111698,
  'Bin 4': 1.1364715170302837}},
'installment_group_Bin': {'iv': 0.028446931068475675,
 'woe': {'Bin 0': -0.2965362458671355,
  'Bin 1': -0.027032933155407588,
  'Bin 2': 0.11545123400239903}},
'employmentTitle_group_Bin': {'iv': 0.01199262426711711,
 'woe': {'Bin 0': 0.1974127496581969,
  'Bin 1': 0.015794613325623683,
  'Bin 2': -0.19061653548410487}},
'homeOwnership_group_Bin': {'iv': 0.0003224362597742428,
 'woe': {'Bin 0': -0.006310542058851011, 'Bin 1': 0.051096232036222884}},
'annualIncome_group_Bin': {'iv': 0.028755712221678248,
 'woe': {'Bin 0': 0.19386504291971357,
  'Bin 1': 0.09133649436575257,
  'Bin 2': -0.026827665423123023,
  'Bin 3': -0.19469768671127013,
  'Bin 4': -0.339204930328438}},
'purpose_group_Bin': {'iv': 0.007295208182839682,
 'woe': {'Bin 0': 0.060104555523002035, 'Bin 1': -0.12144911490483187}},
'postCode_group_Bin': {'iv': 0.0008119510164983021,
 'woe': {'Bin 0': -0.10613844771754728,
  'Bin 1': -0.006970955095481508,
  'Bin 2': 0.02290230125349964}},
'regionCode_group_Bin': {'iv': 1.3326503587742891e-05,
 'woe': {'Bin 0': 0.0037306864251708577, 'Bin 1': -0.003572135757512189}},
'dti_group_Bin': {'iv': 0.0691444505823585,
 'woe': {'Bin 0': -0.29639426975460115,
  'Bin 1': -0.05303861299797819,
  'Bin 2': 0.14116987973826364,
  'Bin 3': 0.2922661709950033,
  'Bin 4': 0.49293116504622203}},
'delinquency_2years_group_Bin': {'iv': 0.002139337579274896,
 'woe': {'Bin 0': -0.012315786189818509,
  'Bin 1': 0.14386566036838583,
  'Bin 2': 0.20305465903034065}},
'ficoRangeLow_group_Bin': {'iv': 0.11849558508830776,
 'woe': {'Bin 0': 0.33499130235102315,
  'Bin 1': 0.14036402817126167,
  'Bin 2': -0.16102580213328785,
  'Bin 3': -0.582451594346089,
  'Bin 4': -1.0396374414819025}},
'ficoRangeHigh_group_Bin': {'iv': 0.11849558508830776,
 'woe': {'Bin 0': 0.33499130235102315,
  'Bin 1': 0.14036402817126167,
  'Bin 2': -0.16102580213328785,
  'Bin 3': -0.582451594346089,
  'Bin 4': -1.0396374414819025}},
'openAcc_group_Bin': {'iv': 0.004573486625874178,
 'woe': {'Bin 0': -0.0890448653651065,
  'Bin 1': -0.030985128792218982,
  'Bin 2': 0.03156043510552112,
  'Bin 3': 0.08247580996942927,
  'Bin 4': 0.18573680489483868}},
'pubRec_group_Bin': {'iv': 0.001300898810070593,
 'woe': {'Bin 0': -0.006265183515565402, 'Bin 1': 0.20766187334039157}},
'pubRecBankruptcies_group_Bin': {'iv': 0.004275157313198444,
 'woe': {'Bin 0': -0.025198889500429856,
  'Bin 1': 0.16242870100858983,
  'Bin 2': 0.2476354559457032}},
'revolBal_group_Bin': {'iv': 0.003766992022044162,
 'woe': {'Bin 0': 0.02404226891591574,
  'Bin 1': -0.05292092268354001,
  'Bin 2': -0.24978853781535482}},
'revolUtil_group_Bin': {'iv': 0.02397767300459437,
 'woe': {'Bin 0': -0.358979455056802,
  'Bin 1': -0.17312304317097293,
  'Bin 2': -0.021333642551765307,
  'Bin 3': 0.09298419081740512,
  'Bin 4': 0.1599708320355574}},
'totalAcc_group_Bin': {'iv': 0.001774565948034374,
 'woe': {'Bin 0': 0.0654454812926487,
  'Bin 1': 0.0046315344580754,
  'Bin 2': -0.03502643818846728}},
'title_group_Bin': {'iv': 0.02138620287983934,
 'woe': {'Bin 0': 0.12076691332729704,
  'Bin 1': -0.06234571640663273,
  'Bin 2': -0.303010040073495}},
'n0_group_Bin': {'iv': 0.0036482088372045843,
 'woe': {'Bin 0': -0.03328497196626471, 'Bin 1': 0.10963861162794214}},
'n1_group_Bin': {'iv': 0.012776531508686802,
 'woe': {'Bin 0': -0.1139367854446227,
  'Bin 1': -0.03145428250504604,
  'Bin 2': 0.029134607317330145,
  'Bin 3': 0.10471618734510002,
  'Bin 4': 0.24003074689416587}},
'n2_group_Bin': {'iv': 0.03305847723307784,
 'woe': {'Bin 0': -0.2599793507860163,
  'Bin 1': -0.1159979609788405,
  'Bin 2': 0.03678331487568212,
  'Bin 3': 0.19935833147539536,
  'Bin 4': 0.3818852049766906}},
'n3_group_Bin': {'iv': 0.03305847723307784,
 'woe': {'Bin 0': -0.2599793507860163,
  'Bin 1': -0.1159979609788405,
  'Bin 2': 0.03678331487568212,
  'Bin 3': 0.19935833147539536,
  'Bin 4': 0.3818852049766906}},
'n4_group_Bin': {'iv': 0.005517283907591153,
 'woe': {'Bin 0': -0.2889421330249432,
  'Bin 1': -0.013202349900056868,
  'Bin 2': 0.019088384219396928,
  'Bin 3': 0.06247232677779756,
  'Bin 4': 0.13402280957203358}},
'n5_group_Bin': {'iv': 0.00020945793402011507,
 'woe': {'Bin 0': 0.023494043351489774, 'Bin 1': -0.008915518979970272}},
'n6_group_Bin': {'iv': 0.0014692680358088994,
 'woe': {'Bin 0': -0.0765558185257892,
  'Bin 1': 0.0053247020248672955,
  'Bin 2': 0.1210950403409298}},
'n7_group_Bin': {'iv': 0.009675687051079466,
 'woe': {'Bin 0': -0.14429755533030605,
  'Bin 1': -0.027727108030571924,
  'Bin 2': 0.041131450162872625,
  'Bin 3': 0.08806082871714123,
  'Bin 4': 0.18795745190250174}},
'n8_group_Bin': {'iv': 0.0007184262698178514,
 'woe': {'Bin 0': -0.06987676998104353, 'Bin 1': 0.010281947543667881}},
'n9_group_Bin': {'iv': 0.031249289588835223,
 'woe': {'Bin 0': -0.2096436342241601,
  'Bin 1': -0.0429081760617028,
  'Bin 2': 0.07105221873661946,
  'Bin 3': 0.1970894208174422,
  'Bin 4': 0.3774005738856138}},
'n10_group_Bin': {'iv': 0.008444830436101544,
 'woe': {'Bin 0': -0.293226157535438,
  'Bin 1': -0.06274924779747353,
  'Bin 2': -0.00328536625477653,
  'Bin 3': 0.05386221291998551,
  'Bin 4': 0.19061556079370015}},
'n11_group_Bin': {'iv': 7.421631321053656e-08,
 'woe': {'Bin 0': 7.237673677632179e-06, 'Bin 1': -0.010254233468312956}},
'n12_group_Bin': {'iv': 7.759942856943026e-05,
 'woe': {'Bin 0': -0.0004960486167429998, 'Bin 1': 0.15643612383430344}},
'n13_group_Bin': {'iv': 0.0015580211455861362,
 'woe': {'Bin 0': -0.009717341239876488, 'Bin 1': 0.1603549088517746}},
'n14_group_Bin': {'iv': 0.048440899304046665,
 'woe': {'Bin 0': -0.3045306871267572,
  'Bin 1': -0.14851272009664643,
  'Bin 2': 0.0023652073777263066,
  'Bin 3': 0.14618635044605022,
  'Bin 4': 0.3310966221715067}}}}

字符变量分组调用

char_group_chi_main(data_total,id_name,y,group_num,drop_value=[])

data_total:原始数据集

id_name:主键名字，例如"idno"

y:y变量名,例如"isbad"

group_num：最优分组组数，一般设定为4-6组

drop_value:不参与最优分组的变量，例如字符型的日期这一类变量。

例如以下填写参数：

char_group_chi_main(train_data,"id","isDefault",5,drop_value=["issueDate","earliesCreditLine"])

返回值：与数值的变量的一样，不再赘述：

{'data':             id  loanAmnt  ...  employmentLength_group  employmentLength_group_Bin
0            0   35000.0  ...                       8                       Bin 2
1            1   18000.0  ...                       4                       Bin 1
2            2   12000.0  ...                       5                       Bin 2
3            3   11000.0  ...                       1                       Bin 0
4            4    3000.0  ...                      12                       Bin 4
      ...       ...  ...                     ...                         ...
799995  799995   25000.0  ...                       3                       Bin 1
799996  799996   17000.0  ...                       1                       Bin 0
799997  799997    6000.0  ...                       1                       Bin 0
799998  799998   19200.0  ...                       1                       Bin 0
799999  799999    9000.0  ...                       4                       Bin 1

[800000 rows x 54 columns],
'var_dict_total': {'grade': {'bin': {'F': 'Bin 4',
   'G': 'Bin 4',
   'D': 'Bin 3',
   'E': 'Bin 4',
   'C': 'Bin 2',
   'A': 'Bin 0',
   'B': 'Bin 1'},
  'var_name': 'grade_group_Bin'},
 'subGrade': {'bin': {'G1': 'Bin 4',
   'G2': 'Bin 4',
   'G3': 'Bin 4',
   'F4': 'Bin 4',
   'F3': 'Bin 4',
   'G4': 'Bin 4',
   'E3': 'Bin 4',
   'F5': 'Bin 4',
   'E4': 'Bin 4',
   'D3': 'Bin 3',
   'D2': 'Bin 3',
   'F1': 'Bin 4',
   'A2': 'Bin 0',
   'F2': 'Bin 4',
   'A5': 'Bin 0',
   'A3': 'Bin 0',
   'B5': 'Bin 1',
   'B4': 'Bin 1',
   'E5': 'Bin 4',
   'G5': 'Bin 4',
   'D1': 'Bin 3',
   'E1': 'Bin 4',
   'C1': 'Bin 2',
   'D4': 'Bin 3',
   'B1': 'Bin 1',
   'D5': 'Bin 3',
   'C2': 'Bin 2',
   'B2': 'Bin 1',
   'C5': 'Bin 3',
   'A1': 'Bin 0',
   'E2': 'Bin 4',
   'C4': 'Bin 3',
   'C3': 'Bin 2',
   'A4': 'Bin 0',
   'B3': 'Bin 1'},
  'var_name': 'subGrade_group_Bin'},
 'employmentLength': {'bin': {'6 years': 'Bin 1',
   '3 years': 'Bin 2',
   '9 years': 'Bin 2',
   '4 years': 'Bin 2',
   '< 1 year': 'Bin 3',
   '8 years': 'Bin 2',
   '1 year': 'Bin 3',
   'null': 'Bin 4',
   '2 years': 'Bin 2',
   '7 years': 'Bin 1',
   '10+ years': 'Bin 0',
   '5 years': 'Bin 1'},
  'var_name': 'employmentLength_group_Bin'}},
'var_bin_list': ['grade_group_Bin',
 'subGrade_group_Bin',
 'employmentLength_group_Bin'],
'iv_dict': {'grade_group_Bin': {'iv': 0.46008926359723423,
  'woe': {'Bin 0': -1.355565869742944,
   'Bin 1': -0.4854115082757125,
   'Bin 2': 0.15269375877204816,
   'Bin 3': 0.560321181232565,
   'Bin 4': 1.0184060522289387}},
 'subGrade_group_Bin': {'iv': 0.46133308496995795,
  'woe': {'Bin 0': -1.355565869742944,
   'Bin 1': -0.4854115082757125,
   'Bin 2': 0.04570655110165618,
   'Bin 3': 0.46384538567636546,
   'Bin 4': 1.0184060522289387}},
 'employmentLength_group_Bin': {'iv': 0.012227612557761728,
  'woe': {'Bin 0': -0.08002342137428835,
   'Bin 1': -0.03072040532440133,
   'Bin 2': -0.00026093587689965793,
   'Bin 3': 0.03412595689559949,
   'Bin 4': 0.38856251491409827}}}}

一个语句建模

刚才的最优分组都是可以单独使用的，一个语句建模，即只需要执行一个语句，则能将原始的数据集从分组、转换woe、筛选变量，拟合模型、最终得到测试集以及训练集的结果。

定义模型参数：

    m=MakeModel(data,"id","isDefault",0.02,5,512254,drop_value=["issueDate","earliesCreditLine"])
    def __init__(self,data,id_name,y_name,iv_num,group_num,random_state,drop_value=[]):

data:原始数据集，最好是你处理过的数据集，具体的衍生变量你自己做啦~~~

id_name：主键的变量名 例如 "idno"

iv_num：设定iv大于多少就可以进入模型拟合 建议是0.01-0.05之间

y_name:y变量名 例如"isbad"

group_num:最优分组最终组数，最好的4-6之间

random_state：划分数据集的随机码

drop_value：不参与建模的变量，你可以自己在原始数据集里面删掉，也可以在这写上。

###执行语句

   m=MakeModel(data,"id","isDefault",0.02,5,512254,drop_value=["issueDate","earliesCreditLine"])
   m.model_main()

最终m包含以下信息： bin_woe 这是类中定义的一个用户转换woe的函数，不管！！

char_object：这个是最终字符变量最优分组输出的信息，上面贴过了，我再贴一次，因为我知道你懒得翻回去

{'data':             id  loanAmnt  ...  employmentLength_group  employmentLength_group_Bin
742562  742562   10000.0  ...                       1                       Bin 0
762848  762848   14950.0  ...                       6                       Bin 2
644240  644240   20000.0  ...                      12                       Bin 4
451072  451072   12000.0  ...                       9                       Bin 3
93411    93411    2500.0  ...                      11                       Bin 3
     ...       ...  ...                     ...                         ...
245245  245245    4450.0  ...                       9                       Bin 3
30730    30730    5700.0  ...                       1                       Bin 0
476922  476922    7500.0  ...                       2                       Bin 1
476992  476992   15000.0  ...                      10                       Bin 3
39730    39730   40000.0  ...                       2                       Bin 1

[560000 rows x 56 columns],
'var_dict_total': {'grade': {'bin': {'G': 'Bin 4',
  'E': 'Bin 4',
  'F': 'Bin 4',
  'D': 'Bin 3',
  'C': 'Bin 2',
  'B': 'Bin 1',
  'A': 'Bin 0'},
 'var_name': 'grade_group_Bin'},
'subGrade': {'bin': {'G5': 'Bin 4',
  'G4': 'Bin 4',
  'F2': 'Bin 4',
  'G3': 'Bin 4',
  'G2': 'Bin 4',
  'F5': 'Bin 4',
  'F4': 'Bin 4',
  'E3': 'Bin 4',
  'G1': 'Bin 4',
  'E4': 'Bin 4',
  'A2': 'Bin 0',
  'A1': 'Bin 0',
  'A3': 'Bin 0',
  'E1': 'Bin 3',
  'D1': 'Bin 3',
  'C4': 'Bin 3',
  'D2': 'Bin 3',
  'E5': 'Bin 4',
  'C3': 'Bin 3',
  'D5': 'Bin 3',
  'E2': 'Bin 4',
  'B4': 'Bin 2',
  'F1': 'Bin 4',
  'C5': 'Bin 3',
  'B1': 'Bin 1',
  'C1': 'Bin 2',
  'D4': 'Bin 3',
  'A4': 'Bin 0',
  'B2': 'Bin 1',
  'F3': 'Bin 4',
  'D3': 'Bin 3',
  'C2': 'Bin 2',
  'B5': 'Bin 2',
  'B3': 'Bin 1',
  'A5': 'Bin 1'},
 'var_name': 'subGrade_group_Bin'},
'employmentLength': {'bin': {'9 years': 'Bin 2',
  '4 years': 'Bin 2',
  '8 years': 'Bin 2',
  '7 years': 'Bin 1',
  '5 years': 'Bin 1',
  'null': 'Bin 4',
  '1 year': 'Bin 3',
  '2 years': 'Bin 2',
  '3 years': 'Bin 3',
  '10+ years': 'Bin 0',
  '< 1 year': 'Bin 3',
  '6 years': 'Bin 1'},
 'var_name': 'employmentLength_group_Bin'}},
'var_bin_list': ['grade_group_Bin',
'subGrade_group_Bin',
'employmentLength_group_Bin'],
'iv_dict': {'grade_group_Bin': {'iv': 0.46497204120261115,
 'woe': {'Bin 0': -1.366114993583636,
  'Bin 1': -0.4872500245227714,
  'Bin 2': 0.15134427847202556,
  'Bin 3': 0.5673843591805174,
  'Bin 4': 1.0199366732850919}},
'subGrade_group_Bin': {'iv': 0.46933032966767474,
 'woe': {'Bin 0': -1.5499290796800491,
  'Bin 1': -0.7155043413125765,
  'Bin 2': -0.14179350218267067,
  'Bin 3': 0.43440570785349064,
  'Bin 4': 1.0693024724239095}},
'employmentLength_group_Bin': {'iv': 0.011729759860744725,
 'woe': {'Bin 0': -0.07860890366978489,
  'Bin 1': -0.030637664745026258,
  'Bin 2': -0.005312943616562383,
  'Bin 3': 0.02768067382373055,
  'Bin 4': 0.3796614757205021}}}}

corr_x:定义的一个计算相关系数矩阵的函数，不管！！！

deal_raw_data：处理映射的函数，不管！！

drop_value： 你输入的删掉的变量的参数

feature_list：进入模型拟合的变量列表，如果你对选出来的变量不满意，可以自己选变量，这样子跑就可以了

 m=MakeModel(data,"id","isDefault",0.02,5,512254,drop_value=["issueDate","earliesCreditLine"])
 m.model_main(feature_list=["aa","bb","dd","cc"])

group_num:你输入的分组数

id_name：你输入的主键的变量名

iv_num：你输入的iv阈值

ks_calc_auc：一个计算ks的函数，不用管。你想用，我之后给你放出来，这个版本不放

model_main：建模主函数，不用管

num_object：这个是最终数值变量最优分组输出的信息，上面贴过了，我再贴一次，因为我知道你懒得翻回去

{'data':             id  loanAmnt  term  ...  n13_group_Bin  n14_group n14_group_Bin
384065  384065    7500.0     3  ...          Bin 0          1         Bin 0
416852  416852   35000.0     5  ...          Bin 0          1         Bin 0
794591  794591   30000.0     5  ...          Bin 0          1         Bin 0
686008  686008    5000.0     3  ...          Bin 0          1         Bin 0
449392  449392   12000.0     3  ...          Bin 0          1         Bin 0
      ...       ...   ...  ...            ...        ...           ...
764543  764543   12000.0     3  ...          Bin 0          7         Bin 4
230131  230131    5550.0     3  ...          Bin 1          7         Bin 4
289950  289950    2200.0     3  ...          Bin 0          7         Bin 4
95878    95878    5500.0     3  ...          Bin 1          7         Bin 4
157628  157628    7200.0     3  ...          Bin 0          7         Bin 4

[560000 rows x 117 columns],
'var_dict_total': {'loanAmnt': {'bin': {'Bin 0': [-9999.0, 8625.0],
   'Bin 1': [8625.0, 10000.0],
   'Bin 2': [10000.0, 15000.0],
   'Bin 3': [15000.0, 28000.0],
   'Bin 4': [28000.0, 10000000000]},
  'value_name': 'loanAmnt_group_Bin'},
 'interestRate': {'bin': {'Bin 0': [-9999.0, 7.97],
   'Bin 1': [7.97, 12.29],
   'Bin 2': [12.29, 15.99],
   'Bin 3': [15.99, 22.15],
   'Bin 4': [22.15, 10000000000]},
  'value_name': 'interestRate_group_Bin'},
 'installment': {'bin': {'Bin 0': [-9999.0, 248.25],
   'Bin 1': [248.25, 324.07],
   'Bin 2': [324.07, 10000000000]},
  'value_name': 'installment_group_Bin'},
 'employmentTitle': {'bin': {'Bin 0': [-9999.0, 54.0],
   'Bin 1': [54.0, 169060.0],
   'Bin 2': [169060.0, 10000000000]},
  'value_name': 'employmentTitle_group_Bin'},
 'homeOwnership': {'bin': {'Bin 0': [-9999.0, 1],
   'Bin 1': [1, 2],
   'Bin 2': [2, 10000000000]},
  'value_name': 'homeOwnership_group_Bin'},
 'annualIncome': {'bin': {'Bin 0': [-9999.0, 42000.0],
   'Bin 1': [42000.0, 60000.0],
   'Bin 2': [60000.0, 85000.0],
   'Bin 3': [85000.0, 125000.0],
   'Bin 4': [125000.0, 10000000000]},
  'value_name': 'annualIncome_group_Bin'},
 'purpose': {'bin': {'Bin 0': [-9999.0, 2], 'Bin 1': [2, 10000000000]},
  'value_name': 'purpose_group_Bin'},
 'postCode': {'bin': {'Bin 0': [-9999.0, 19.0],
   'Bin 1': [19.0, 300.0],
   'Bin 2': [300.0, 10000000000]},
  'value_name': 'postCode_group_Bin'},
 'regionCode': {'bin': {'Bin 0': [-9999.0, 21], 'Bin 1': [21, 10000000000]},
  'value_name': 'regionCode_group_Bin'},
 'dti': {'bin': {'Bin 0': [-9999.0, 15.31],
   'Bin 1': [15.31, 21.25],
   'Bin 2': [21.25, 25.69],
   'Bin 3': [25.69, 29.78],
   'Bin 4': [29.78, 10000000000]},
  'value_name': 'dti_group_Bin'},
 'delinquency_2years': {'bin': {'Bin 0': [-9999.0, 1.0],
   'Bin 1': [1.0, 2.0],
   'Bin 2': [2.0, 10000000000]},
  'value_name': 'delinquency_2years_group_Bin'},
 'ficoRangeLow': {'bin': {'Bin 0': [-9999.0, 680.0],
   'Bin 1': [680.0, 705.0],
   'Bin 2': [705.0, 720.0],
   'Bin 3': [720.0, 760.0],
   'Bin 4': [760.0, 10000000000]},
  'value_name': 'ficoRangeLow_group_Bin'},
 'ficoRangeHigh': {'bin': {'Bin 0': [-9999.0, 684.0],
   'Bin 1': [684.0, 709.0],
   'Bin 2': [709.0, 724.0],
   'Bin 3': [724.0, 764.0],
   'Bin 4': [764.0, 10000000000]},
  'value_name': 'ficoRangeHigh_group_Bin'},
 'openAcc': {'bin': {'Bin 0': [-9999.0, 8.0],
   'Bin 1': [8.0, 10.0],
   'Bin 2': [10.0, 19.0],
   'Bin 3': [19.0, 22.0],
   'Bin 4': [22.0, 10000000000]},
  'value_name': 'openAcc_group_Bin'},
 'pubRec': {'bin': {'Bin 0': [-9999.0, 1.0], 'Bin 1': [1.0, 10000000000]},
  'value_name': 'pubRec_group_Bin'},
 'pubRecBankruptcies': {'bin': {'Bin 0': [-9999.0, 0.0],
   'Bin 1': [0.0, 1.0],
   'Bin 2': [1.0, 10000000000]},
  'value_name': 'pubRecBankruptcies_group_Bin'},
 'revolBal': {'bin': {'Bin 0': [-9999.0, 22687.0],
   'Bin 1': [22687.0, 43329.0],
   'Bin 2': [43329.0, 10000000000]},
  'value_name': 'revolBal_group_Bin'},
 'revolUtil': {'bin': {'Bin 0': [-9999.0, 18.1],
   'Bin 1': [18.1, 37.4],
   'Bin 2': [37.4, 55.6],
   'Bin 3': [55.6, 74.9],
   'Bin 4': [74.9, 10000000000]},
  'value_name': 'revolUtil_group_Bin'},
 'totalAcc': {'bin': {'Bin 0': [-9999.0, 16.0],
   'Bin 1': [16.0, 22.0],
   'Bin 2': [22.0, 26.0],
   'Bin 3': [26.0, 10000000000]},
  'value_name': 'totalAcc_group_Bin'},
 'title': {'bin': {'Bin 0': [-9999.0, 1.0],
   'Bin 1': [1.0, 10.0],
   'Bin 2': [10.0, 10000000000]},
  'value_name': 'title_group_Bin'},
 'n0': {'bin': {'Bin 0': [-9999.0, 0.0], 'Bin 1': [0.0, 10000000000]},
  'value_name': 'n0_group_Bin'},
 'n1': {'bin': {'Bin 0': [-9999.0, 2.0],
   'Bin 1': [2.0, 3.0],
   'Bin 2': [3.0, 4.0],
   'Bin 3': [4.0, 6.0],
   'Bin 4': [6.0, 10000000000]},
  'value_name': 'n1_group_Bin'},
 'n2': {'bin': {'Bin 0': [-9999.0, 2.0],
   'Bin 1': [2.0, 4.0],
   'Bin 2': [4.0, 6.0],
   'Bin 3': [6.0, 8.0],
   'Bin 4': [8.0, 10000000000]},
  'value_name': 'n2_group_Bin'},
 'n3': {'bin': {'Bin 0': [-9999.0, 2.0],
   'Bin 1': [2.0, 4.0],
   'Bin 2': [4.0, 6.0],
   'Bin 3': [6.0, 8.0],
   'Bin 4': [8.0, 10000000000]},
  'value_name': 'n3_group_Bin'},
 'n4': {'bin': {'Bin 0': [-9999.0, 0.0],
   'Bin 1': [0.0, 4.0],
   'Bin 2': [4.0, 6.0],
   'Bin 3': [6.0, 10000000000]},
  'value_name': 'n4_group_Bin'},
 'n5': {'bin': {'Bin 0': [-9999.0, 4.0], 'Bin 1': [4.0, 10000000000]},
  'value_name': 'n5_group_Bin'},
 'n6': {'bin': {'Bin 0': [-9999.0, 1.0],
   'Bin 1': [1.0, 23.0],
   'Bin 2': [23.0, 10000000000]},
  'value_name': 'n6_group_Bin'},
 'n7': {'bin': {'Bin 0': [-9999.0, 4.0],
   'Bin 1': [4.0, 7.0],
   'Bin 2': [7.0, 10.0],
   'Bin 3': [10.0, 14.0],
   'Bin 4': [14.0, 10000000000]},
  'value_name': 'n7_group_Bin'},
 'n8': {'bin': {'Bin 0': [-9999.0, 5.0], 'Bin 1': [5.0, 10000000000]},
  'value_name': 'n8_group_Bin'},
 'n9': {'bin': {'Bin 0': [-9999.0, 3.0],
   'Bin 1': [3.0, 5.0],
   'Bin 2': [5.0, 7.0],
   'Bin 3': [7.0, 10.0],
   'Bin 4': [10.0, 10000000000]},
  'value_name': 'n9_group_Bin'},
 'n10': {'bin': {'Bin 0': [-9999.0, 3.0],
   'Bin 1': [3.0, 8.0],
   'Bin 2': [8.0, 10.0],
   'Bin 3': [10.0, 19.0],
   'Bin 4': [19.0, 10000000000]},
  'value_name': 'n10_group_Bin'},
 'n12': {'bin': {'Bin 0': [-9999.0, 0.0], 'Bin 1': [0.0, 10000000000]},
  'value_name': 'n12_group_Bin'},
 'n13': {'bin': {'Bin 0': [-9999.0, 0.0], 'Bin 1': [0.0, 10000000000]},
  'value_name': 'n13_group_Bin'},
 'n14': {'bin': {'Bin 0': [-9999.0, 0.0],
   'Bin 1': [0.0, 1.0],
   'Bin 2': [1.0, 2.0],
   'Bin 3': [2.0, 3.0],
   'Bin 4': [3.0, 10000000000]},
  'value_name': 'n14_group_Bin'}},
'var_bin_list': ['loanAmnt_group_Bin',
 'interestRate_group_Bin',
 'installment_group_Bin',
 'employmentTitle_group_Bin',
 'homeOwnership_group_Bin',
 'annualIncome_group_Bin',
 'purpose_group_Bin',
 'postCode_group_Bin',
 'regionCode_group_Bin',
 'dti_group_Bin',
 'delinquency_2years_group_Bin',
 'ficoRangeLow_group_Bin',
 'ficoRangeHigh_group_Bin',
 'openAcc_group_Bin',
 'pubRec_group_Bin',
 'pubRecBankruptcies_group_Bin',
 'revolBal_group_Bin',
 'revolUtil_group_Bin',
 'totalAcc_group_Bin',
 'title_group_Bin',
 'n0_group_Bin',
 'n1_group_Bin',
 'n2_group_Bin',
 'n3_group_Bin',
 'n4_group_Bin',
 'n5_group_Bin',
 'n6_group_Bin',
 'n7_group_Bin',
 'n8_group_Bin',
 'n9_group_Bin',
 'n10_group_Bin',
 'n12_group_Bin',
 'n13_group_Bin',
 'n14_group_Bin',
 'term',
 'verificationStatus',
 'initialListStatus',
 'applicationType',
 'policyCode',
 'n11'],
'iv_dict': {'term': {'iv': 0.17475272449362428,
  'woe': {3: -0.2703191679368651, 5: 0.6559694111951718}},
 'verificationStatus': {'iv': 0.053778743185793496,
  'woe': {0: -0.36484279075332327,
   1: 0.06141693489701235,
   2: 0.22266031298696004}},
 'initialListStatus': {'iv': 0.00038571827775202977,
  'woe': {0: 0.01649700561049484, 1: -0.023381860067453963}},
 'applicationType': {'iv': 0.001984746493023151,
  'woe': {0: -0.006543082114183424, 1: 0.3033852042307132}},
 'policyCode': {'iv': 0.0, 'woe': {1.0: 0.0}},
 'n11': {'iv': 5.669110692040425e-05,
  'woe': {0.0: -0.0013462832120869654,
   1.0: 0.0003796115927952374,
   2.0: -1.321366999272626,
   4.0: -10.126261596374201}},
 'loanAmnt_group_Bin': {'iv': 0.03343117392609739,
  'woe': {'Bin 0': -0.25565091759789366,
   'Bin 1': -0.10181829737829946,
   'Bin 2': 0.048523004375282064,
   'Bin 3': 0.15719425693473638,
   'Bin 4': 0.2554982959104215}},
 'interestRate_group_Bin': {'iv': 0.43944475156239415,
  'woe': {'Bin 0': -1.4336529379346452,
   'Bin 1': -0.47629575281922726,
   'Bin 2': 0.1260287233070282,
   'Bin 3': 0.638424679929549,
   'Bin 4': 1.1369940841694457}},
 'installment_group_Bin': {'iv': 0.027688805426894836,
  'woe': {'Bin 0': -0.29335432895221053,
   'Bin 1': -0.02248530913230762,
   'Bin 2': 0.11333264072431821}},
 'employmentTitle_group_Bin': {'iv': 0.011673058473026044,
  'woe': {'Bin 0': 0.19241479983416104,
   'Bin 1': 0.01650461957193068,
   'Bin 2': -0.18949121362830715}},
 'homeOwnership_group_Bin': {'iv': 0.0002916379612195541,
  'woe': {'Bin 0': -0.00598818429937241,
   'Bin 1': 0.04866764184572045,
   'Bin 2': 0.0582736696198456}},
 'annualIncome_group_Bin': {'iv': 0.028734272942132155,
  'woe': {'Bin 0': 0.20550314786386936,
   'Bin 1': 0.1013731532894062,
   'Bin 2': -0.029439154506365764,
   'Bin 3': -0.19611334622465693,
   'Bin 4': -0.3341722425856623}},
 'purpose_group_Bin': {'iv': 0.007398348262121109,
  'woe': {'Bin 0': 0.06057157985102101, 'Bin 1': -0.12221757368584851}},
 'postCode_group_Bin': {'iv': 0.0007570488920220754,
  'woe': {'Bin 0': -0.10132004266079044,
   'Bin 1': -0.005811282866672897,
   'Bin 2': 0.02460536621769406}},
 'regionCode_group_Bin': {'iv': 0.0003170244903853317,
  'woe': {'Bin 0': 0.010672734282877688, 'Bin 1': -0.029704933969456917}},
 'dti_group_Bin': {'iv': 0.06736335316502741,
  'woe': {'Bin 0': -0.2732685235815132,
   'Bin 1': -0.03476878794313418,
   'Bin 2': 0.13763524255784043,
   'Bin 3': 0.28987300233823476,
   'Bin 4': 0.4903505432780388}},
 'delinquency_2years_group_Bin': {'iv': 0.0022623712300652897,
  'woe': {'Bin 0': -0.012708363930036529,
   'Bin 1': 0.1475062082023903,
   'Bin 2': 0.20801887412532466}},
 'ficoRangeLow_group_Bin': {'iv': 0.12055544422234395,
  'woe': {'Bin 0': 0.28837473031500915,
   'Bin 1': -0.008500511432484521,
   'Bin 2': -0.30104870120279137,
   'Bin 3': -0.5875041637976371,
   'Bin 4': -1.0326430545002714}},
 'ficoRangeHigh_group_Bin': {'iv': 0.12055544422234395,
  'woe': {'Bin 0': 0.28837473031500915,
   'Bin 1': -0.008500511432484521,
   'Bin 2': -0.30104870120279137,
   'Bin 3': -0.5875041637976371,
   'Bin 4': -1.0326430545002714}},
 'openAcc_group_Bin': {'iv': 0.004306029382523432,
  'woe': {'Bin 0': -0.07821428563153432,
   'Bin 1': -0.019151061460234813,
   'Bin 2': 0.034864790449546074,
   'Bin 3': 0.10529612262224165,
   'Bin 4': 0.17443385861739738}},
 'pubRec_group_Bin': {'iv': 0.0012341393260821526,
  'woe': {'Bin 0': -0.0061018199179689245, 'Bin 1': 0.20227836670835725}},
 'pubRecBankruptcies_group_Bin': {'iv': 0.004386867849714986,
  'woe': {'Bin 0': -0.025551658775713103,
   'Bin 1': 0.16462934316291766,
   'Bin 2': 0.2489311629377492}},
 'revolBal_group_Bin': {'iv': 0.003941432411932272,
  'woe': {'Bin 0': 0.025043463019543088,
   'Bin 1': -0.057776412254312,
   'Bin 2': -0.2522823282700653}},
 'revolUtil_group_Bin': {'iv': 0.02315807262983133,
  'woe': {'Bin 0': -0.34972229902847746,
   'Bin 1': -0.15293900491846227,
   'Bin 2': 0.012278839217558546,
   'Bin 3': 0.09393480729600885,
   'Bin 4': 0.16029349283335198}},
 'totalAcc_group_Bin': {'iv': 0.001898851737310343,
  'woe': {'Bin 0': 0.06652798510153149,
   'Bin 1': 0.0063347829562498655,
   'Bin 2': -0.022235005488180352,
   'Bin 3': -0.04113034955049715}},
 'title_group_Bin': {'iv': 0.021017346706830586,
  'woe': {'Bin 0': 0.12026055634191922,
   'Bin 1': -0.06297069710648713,
   'Bin 2': -0.29878201272196026}},
 'n0_group_Bin': {'iv': 0.0038659161878576197,
  'woe': {'Bin 0': -0.03432653353796549, 'Bin 1': 0.11265809331135226}},
 'n1_group_Bin': {'iv': 0.01251701221843841,
  'woe': {'Bin 0': -0.11118731349483088,
   'Bin 1': -0.030385344572322486,
   'Bin 2': 0.024394768158194042,
   'Bin 3': 0.10070490376937744,
   'Bin 4': 0.24324257832089885}},
 'n2_group_Bin': {'iv': 0.03182485856762544,
  'woe': {'Bin 0': -0.25234529542718825,
   'Bin 1': -0.12039210786755326,
   'Bin 2': 0.01654240445119373,
   'Bin 3': 0.12840398734442102,
   'Bin 4': 0.30175137102433}},
 'n3_group_Bin': {'iv': 0.03182485856762544,
  'woe': {'Bin 0': -0.25234529542718825,
   'Bin 1': -0.12039210786755326,
   'Bin 2': 0.01654240445119373,
   'Bin 3': 0.12840398734442102,
   'Bin 4': 0.30175137102433}},
 'n4_group_Bin': {'iv': 0.005267251015654909,
  'woe': {'Bin 0': -0.28953419649589196,
   'Bin 1': -0.011632803273824087,
   'Bin 2': 0.016095261458987927,
   'Bin 3': 0.07691228812277608}},
 'n5_group_Bin': {'iv': 0.0002521958240056829,
  'woe': {'Bin 0': 0.025796860662542026, 'Bin 1': -0.009776426926970056}},
 'n6_group_Bin': {'iv': 0.0013155150687629844,
  'woe': {'Bin 0': -0.07315867586330163,
   'Bin 1': 0.005254489185691692,
   'Bin 2': 0.11321926249888056}},
 'n7_group_Bin': {'iv': 0.009545069117979268,
  'woe': {'Bin 0': -0.14454735339316377,
   'Bin 1': -0.024874826581771457,
   'Bin 2': 0.03959338275756879,
   'Bin 3': 0.08367884371104235,
   'Bin 4': 0.18950818931946098}},
 'n8_group_Bin': {'iv': 0.0006247355335945231,
  'woe': {'Bin 0': -0.06522263650073688, 'Bin 1': 0.009579006538587252}},
 'n9_group_Bin': {'iv': 0.030937050220452105,
  'woe': {'Bin 0': -0.20889047134965136,
   'Bin 1': -0.04280319854690665,
   'Bin 2': 0.07380942227136413,
   'Bin 3': 0.19161748950791213,
   'Bin 4': 0.3782251363238346}},
 'n10_group_Bin': {'iv': 0.008253500739303452,
  'woe': {'Bin 0': -0.29328765643532173,
   'Bin 1': -0.06153347064891405,
   'Bin 2': -0.0017968871206732596,
   'Bin 3': 0.047852298082086936,
   'Bin 4': 0.1469804159216567}},
 'n12_group_Bin': {'iv': 8.313258112458898e-05,
  'woe': {'Bin 0': -0.0005117920937828983, 'Bin 1': 0.16243538529117849}},
 'n13_group_Bin': {'iv': 0.0017296224617373412,
  'woe': {'Bin 0': -0.010273646500446446, 'Bin 1': 0.16837952088281963}},
 'n14_group_Bin': {'iv': 0.05013255836162971,
  'woe': {'Bin 0': -0.31114470867320426,
   'Bin 1': -0.15108234506634566,
   'Bin 2': 0.0032456128370830016,
   'Bin 3': 0.15067931120695519,
   'Bin 4': 0.33510301297940864}}}}

result：最终的结果，长这个样子：

{'model_summary': <class 'statsmodels.iolib.summary.Summary'>
"""
                         Logit Regression Results                           
==============================================================================
Dep. Variable:              isDefault   No. Observations:               560000
Model:                          Logit   Df Residuals:                   559984
Method:                           MLE   Df Model:                           15
Date:                Thu, 04 Mar 2021   Pseudo R-squ.:                 0.08853
Time:                        14:28:07   Log-Likelihood:            -2.5537e+05
converged:                       True   LL-Null:                   -2.8018e+05
Covariance Type:            nonrobust   LLR p-value:                     0.000
===========================================================================================
                            coef    std err          z      P>|z|      [0.025      0.975]
-------------------------------------------------------------------------------------------
term                        0.4856      0.010     48.088      0.000       0.466       0.505
verificationStatus          0.2431      0.016     15.175      0.000       0.212       0.275
loanAmnt_group_Bin          0.2776      0.041      6.815      0.000       0.198       0.357
interestRate_group_Bin      0.1153      0.016      7.022      0.000       0.083       0.148
installment_group_Bin       0.2646      0.039      6.728      0.000       0.187       0.342
annualIncome_group_Bin      1.0481      0.025     42.018      0.000       0.999       1.097
dti_group_Bin               0.3618      0.014     25.125      0.000       0.334       0.390
ficoRangeLow_group_Bin      0.1894        nan        nan        nan         nan         nan
ficoRangeHigh_group_Bin     0.1894        nan        nan        nan         nan         nan
revolUtil_group_Bin        -0.2620      0.028     -9.307      0.000      -0.317      -0.207
title_group_Bin             0.4616      0.025     18.401      0.000       0.412       0.511
n2_group_Bin                0.0472        nan        nan        nan         nan         nan
n3_group_Bin                0.0472        nan        nan        nan         nan         nan
n9_group_Bin                0.2948      0.004     74.254      0.000       0.287       0.303
n14_group_Bin               0.3586      0.018     20.174      0.000       0.324       0.393
grade_group_Bin             0.3026      0.017     17.876      0.000       0.269       0.336
subGrade_group_Bin          0.3198      0.014     22.258      0.000       0.292       0.348
intercept                  -1.3887      0.004   -387.449      0.000      -1.396      -1.382
===========================================================================================
""",
'ks': 0.29957454898668295,
'auc': 0.7067054639204269,
'val_ks': 0.29633281875616696,
'val_auc': 0.7043781162452405}

select_value_iv：一个函数来的，不用管啦

train_data：训练集

val_data：测试集

y_name：你输入的y值的变量名

写到我手都废了，你觉得我辛苦的话，可以关注我的公众号：屁屁和铭仔的数据之路

好气哦，不知道怎么贴收款码，万一你们想打赏我们两个怎么办！！！！！

About

评分卡 逻辑回归 最优分组 金融风控 信用等级 屁屁和铭仔的数据之路 AugustBetty

Resources

Readme
Activity

Stars

11 stars

Watchers

1 watching

Forks

3 forks
Report repository

Releases

No releases published

Packages

No packages published

Languages