Get 20M+ Full-Text Papers For Less Than $1.50/day. Start a 14-Day Trial for You or Your Team.

Learn More →

Reply to the paper ‘Do not adjust coefficients in Shapley value regression’ by U. Gromping, S. Landau, Applied Stochastic Models in Business and Industry, 2009; DOI: 10.1002/asmb.773

Reply to the paper ‘Do not adjust coefficients in Shapley value regression’ by U. Gromping, S.... by U. Gromping, S. Landau, Applied Stochastic Models in Business and Industry, 2009; DOI: 10.1002/asmb.773 From: Stan Lipovetsky GfK Custom Research North America, 8401 Golden Valley Rd, Minneapolis, MN 55427, U.S.A. W. Michael Conklin MarketTools, Inc., 6465 Wayzata Blvd, Suite 170, St. Louis Park, MN 55426, U.S.A. Our 2001 paper on Shapley value (SV) regression [1] considered estimation of the individual predictors’ contribution to a model and the re-evaluation of the regression coefficients. A recent contrasting paper [2] claims that only the first part of the estimate of the regressors’ importance is useful, not the re-estimation of the regression coefficients. We thank the authors [2] for attracting attention of statisticians and practitioners to this problem, however, we would like to emphasize that SV regression technique with the adjusted coefficients produces a meaningful model, is helpful for interpretation, and yields efficient predictions. We have been considering different approaches to produce meaningful coefficients for regressions with collinear predictors [3–4]. In a recent supporting work [5] (Tables 2–3 and 6), several models of ordinary least squares (OLS), stepwise OLS, adjusted SV, two-parameter ridge, and restricted regressions expressed via exponential, logit, and multinomial parameterization of the coefficients of linear regression have been http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Applied Stochastic Models in Business and Industry Wiley

Reply to the paper ‘Do not adjust coefficients in Shapley value regression’ by U. Gromping, S. Landau, Applied Stochastic Models in Business and Industry, 2009; DOI: 10.1002/asmb.773

Download PDF
2 pages

Loading next page...
 
/lp/wiley/reply-to-the-paper-do-not-adjust-coefficients-in-shapley-value-SOH2LlnDOR

References (5)

Publisher
Wiley
Copyright
Copyright © 2010 John Wiley & Sons, Ltd.
ISSN
1524-1904
eISSN
1526-4025
DOI
10.1002/asmb.829
Publisher site
See Article on Publisher Site

Abstract

by U. Gromping, S. Landau, Applied Stochastic Models in Business and Industry, 2009; DOI: 10.1002/asmb.773 From: Stan Lipovetsky GfK Custom Research North America, 8401 Golden Valley Rd, Minneapolis, MN 55427, U.S.A. W. Michael Conklin MarketTools, Inc., 6465 Wayzata Blvd, Suite 170, St. Louis Park, MN 55426, U.S.A. Our 2001 paper on Shapley value (SV) regression [1] considered estimation of the individual predictors’ contribution to a model and the re-evaluation of the regression coefficients. A recent contrasting paper [2] claims that only the first part of the estimate of the regressors’ importance is useful, not the re-estimation of the regression coefficients. We thank the authors [2] for attracting attention of statisticians and practitioners to this problem, however, we would like to emphasize that SV regression technique with the adjusted coefficients produces a meaningful model, is helpful for interpretation, and yields efficient predictions. We have been considering different approaches to produce meaningful coefficients for regressions with collinear predictors [3–4]. In a recent supporting work [5] (Tables 2–3 and 6), several models of ordinary least squares (OLS), stepwise OLS, adjusted SV, two-parameter ridge, and restricted regressions expressed via exponential, logit, and multinomial parameterization of the coefficients of linear regression have been

Journal

Applied Stochastic Models in Business and Industry – Wiley

Published: Mar 1, 2010

There are no references for this article.