Altair slc chapter III verifying training logistic using holdout

Roger DeAngelis_21630

Altair slc chapter III verifying training logistic using holdout

Too long to pos on a listserve, see github

github
https://github.com/rogerjdeangelis/utl-altair-slc-chapter-III-verifying-training-logistic-using-holdout

First 2 chapters
https://github.com/rogerjdeangelis/utl-altair-slc-chapter-I-optimum-binning-in-preparation-for-logistic-regression
https://github.com/rogerjdeangelis/utl-altair-slc-chapter-II-identifying-the-best-five-logistic-models

github verification lify cpmparison
https://github.com/rogerjdeangelis/utl-altair-slc-chapter-III-verifying-training-logistic-using-holdout/blob/main/liftplots.png

CONTENTS

 1  Best Model from Chater II

    RESPONSE =_MARRIED _ACSGENDER _DIVISIONCDE _INCOME _LOANHOME


 2  Training Logistic


 3  Create SAS code to bin variables

    Basically, If then else code that will bin the Holdout sample.

    Code is created using the binning done by the optimalbin macro on the training data.


 4  Map char vars  (create sas if then code)


 5  Map num vars   (create sas if then code)


 6  Applly if then elses to holdout to bin data


 7  Run logistic on holdout sample  \


 8  Rank probabilities and modeled probablities


 9  Verify training logistic using holdout


10  Compare Lift plots Training and Holdout


11  Ascii plot verification of Training and Holdout


12  Output datasets

Roger DeAngelis_21630

FIXING AN ISSUE WITH THE ALGORITHM. NEED TO USE THE MODEL EQUATION INSTEAD OF THE HOLDOUT LOGISTIC.