JSM 2014 Home
Online Program Home
My Program

Abstract Details

Activity Number: 167
Type: Contributed
Date/Time: Monday, August 4, 2014 : 10:30 AM to 12:20 PM
Sponsor: Section for Statistical Programmers and Analysts
Abstract #311359 View Presentation
Title: Screening Predictors for Logistic Regression Using the Information Value Statistic
Author(s): Bruce Lund*+
Companies: Marketing Associates
Keywords: Logistic Regression ; Information Value Statistic ; Optimal Binning of Predictors ; Direct Marketing
Abstract:

In preparing a nominal or discrete-valued predictor variable X for binary logistic regression it is good practice to collapse the levels of X to achieve parsimony while still maintaining predictive power. In section #1 an algorithm is given for collapsing the levels of X. If ordinal, the ordering of X can be maintained by collapsing only adjacent levels of X. Otherwise, all pairs of levels are considered for collapsing. There is a choice of two criteria for collapsing: maximizing (a) information value (IV) or (b) log likelihood (LL). Stopping guidelines are provided. An example is given where the algorithm is used to create interactions of two predictors. In section #2 the IV statistic is discussed. Familiar, but perhaps mysterious, guidelines for deciding if IV of X is high enough to use in a model are given in textbooks. To provide insight into IV guidelines, the IV is compared to other measures of predictive power. In section #3 a simulation is given of the probability distribution of IV under assumption of no relationship between X and the binary target. Comparison is made to IV guidelines. The focus is on direct marketing and credit score applications.


Authors who are presenting talks have a * after their name.

Back to the full JSM 2014 program




2014 JSM Online Program Home

For information, contact jsm@amstat.org or phone (888) 231-3473.

If you have questions about the Professional Development program, please contact the Education Department.

The views expressed here are those of the individual authors and not necessarily those of the JSM sponsors, their officers, or their staff.

ASA Meetings Department  •  732 North Washington Street, Alexandria, VA 22314  •  (703) 684-1221  •  meetings@amstat.org
Copyright © American Statistical Association.