Model Selection and Estimation in Regression with Grouped Variables

Выбор модели и оценивание в регрессии с группированными переменными
Ming Yuan, Yi Lin
2005-12-21

LARS algorithmanalysis of variancegrouped variable selectionlassonon-negative garrotte
Summary We consider the problem of selecting grouped variables (factors) for accurate prediction in regression. Such a problem arises naturally in many practical situations with the multifactor analysis-of-variance problem as the most important and well-known example. Instead of selecting factors by stepwise backward elimination, we focus on the accuracy of estimation and consider extensions of the lasso, the LARS algorithm and the non-negative garrotte for factor selection. The lasso, the LARS algorithm and the non-negative garrotte are recently proposed regression methods that can be used to select individual variables. We study and propose efficient algorithms for the extensions of these methods for factor selection and show that these extensions give superior performance to the traditional stepwise backward elimination method in factor selection problems. We study the similarities and the differences between these methods. Simulations and real examples are used to illustrate the methods.
1
Extensions of the lasso, LARS algorithm, and non-negative garrote can be formulated to select grouped variables (factors) in regression.
2
Proposed algorithms for these grouped-variable extensions are computationally efficient.
3
The grouped-variable extensions outperform traditional stepwise backward elimination in factor selection problems, based on simulations and real examples.
4
The paper analyzes similarities and differences among the grouped-variable lasso, grouped LARS, and grouped non-negative garrote methods.

Regression models with grouped variables (factors) for factor selection

Accuracy of model selection and estimation for grouped-variable (factor) selection using extensions of the lasso, LARS, and non-negative garrotte compared to stepwise backward elimination

Publication Details
Publication Date
2005-12-21
Journal
Publisher
ISSN
Cited by
7267
Access Type
Author Information
Authors
Ming Yuan
Yi Lin
Explore further
Open the scid.ai AI chat with a ready-made request: it will find papers on a similar topic and help build a literature review.
Find similar papers in the chat →
Make a presentation
100%