Abstract #300322

This is the preliminary program for the 2003 Joint Statistical Meetings in San Francisco, California. Currently included in this program is the "technical" program, schedule of invited, topic contributed, regular contributed and poster sessions; Continuing Education courses (August 2-5, 2003); and Committee and Business Meetings. This on-line program will be updated frequently to reflect the most current revisions.

To View the Program:
You may choose to view all activities of the program or just parts of it at any one time. All activities are arranged by date and time.

The views expressed here are those of the individual authors
and not necessarily those of the ASA or its board, officers, or staff.

Back to main JSM 2003 Program page
JSM 2003 Abstract #300322
Activity Number: 262
Type: Luncheons
Date/Time: Tuesday, August 5, 2003 : 12:30 PM to 1:50 PM
Sponsor: Section on Statistical Consulting
Abstract - #300322
Title: Credit Scoring--Palatable Algorithmic Modeling - THIS LUNCHEON IS SOLD OUT
Author(s): Bruce Hoadley*+
Companies: Fair Isaac Corporation
Address: 2921 Regent St., Berkeley, CA, 94705-2101,
Keywords: credit scoring ; algorithmic modeling ; regression trees ; GAM ; quadratic programming ; logistic regression
Abstract:

In a recent Statistical Science paper, Leo Breiman discusses two cultures of statistical modeling: data modeling and algorithmic modeling. Traditional statistics focuses on data modeling, but modern classification and prediction applications tend to use algorithmic modeling. Breiman argues that you cannot develop algorithmic models that are both interpretable and very accurate. Participants will discuss various aspects of this claim. Credit scoring, as practiced by Fair, Isaac and Company, will be offered as one counterexample to this claim. This is a parallel universe of statistical technology developed somewhat independently of formal data modeling ideas. We hope participants will bring their own evidence--one way or the other. Part of the discussion will be an argument that you can develop excellent algorithmic models with data modeling tools as long as you: (i) ignore most textbook advice; (ii) embrace the blessing of dimensionality; (iii) use constraints in the fitting optimizations; (iv) use regularization; and (v) validate the results.

  • The address information is for the authors that have a + after their name.
  • Authors who are presenting talks have a * after their name.

Back to the full JSM 2003 program

JSM 2003 For information, contact meetings@amstat.org or phone (703) 684-1221. If you have questions about the Continuing Education program, please contact the Education Department.
Revised March 2003