|
Activity Number:
|
464
|
|
Type:
|
Topic Contributed
|
|
Date/Time:
|
Wednesday, August 1, 2007 : 2:00 PM to 3:50 PM
|
|
Sponsor:
|
Section on Physical and Engineering Sciences
|
| Abstract - #309475 |
|
Title:
|
Recent Directions in Predictor Selection
|
|
Author(s):
|
Jessica Kraker*+ and Douglas M. Hawkins
|
|
Companies:
|
University of Wisconsin-Eau Claire and The University of Minnesota
|
|
Address:
|
216 Helen Street, Roberts, WI, 54023,
|
|
Keywords:
|
predictor selection ; penalized regression ; chemometrics
|
|
Abstract:
|
Beginning with the closed-form ridge regression model (quadratic loss and penalty) and advancing to more computationally-intensive methods (such as the lasso and elastic net), the possibilities for penalized regression have progressed dramatically in recent years. In the context of chemometrics, we analyze a prediction problem calling for the concurrent selection of predictors with fitting of the regression model. Model selection from among several penalized regression models (with different loss and penalty functions) requires the further assessment of the model utility. Models are included with consideration for the type of predictors implemented by the researcher and for the type of loss function desired. Programming is implemented in the R environment to obtain and to assess the fitted models.
|
