Variable selection via RIVAL (removing irrelevant variables amidst Lasso iterations) and its application to nuclear material detection

Paul Kump; Er-Wei Bai; Kung-sik Chan; Bill Eichinger; Kang Li

doi:10.1016/j.automatica.2012.06.051

Back

Variable selection via RIVAL (removing irrelevant variables amidst Lasso iterations) and its application to nuclear material detection

Journal article

Peer reviewed

Variable selection via RIVAL (removing irrelevant variables amidst Lasso iterations) and its application to nuclear material detection

Paul Kump, Er-Wei Bai, Kung-sik Chan, Bill Eichinger and Kang Li

Automatica (Oxford), Vol.48(9), pp.2107-2115

09/2012

DOI: 10.1016/j.automatica.2012.06.051

View Online

Abstract

In many situations, the number of data points is fixed, and the asymptotic convergence results of popular model selection tools may not be useful. A new algorithm for model selection, RIVAL (removing irrelevant variables amidst Lasso iterations), is presented and shown to be particularly effective for a large but fixed number of data points. The algorithm is motivated by an application of nuclear material detection where all unknown parameters are to be non-negative. Thus, positive Lasso and its variants are analyzed. Then, RIVAL is proposed and is shown to have some desirable properties, namely the number of data points needed to have convergence is smaller than existing methods.

Positive Lasso

Nuclear material detection application

RIVAL

Model selection

Details

Title: Subtitle: Variable selection via RIVAL (removing irrelevant variables amidst Lasso iterations) and its application to nuclear material detection
Creators: Paul Kump - Department of Electrical and Computer Engineering, University of Iowa, Iowa City, IA 52242, United States
Er-Wei Bai - Department of Electrical and Computer Engineering, University of Iowa, Iowa City, IA 52242, United States
Kung-sik Chan - Department of Statistics, University of Iowa, Iowa City, IA 52242, United States
Bill Eichinger - Department of Civil Engineering, University of Iowa, Iowa City, IA 52242, United States
Kang Li - School of Electronics, Electrical Engineering and Computer Science, Queen’s University, Belfast BT7 1NN, UK
Resource Type: Journal article
Publication Details: Automatica (Oxford), Vol.48(9), pp.2107-2115
Publisher: Elsevier Ltd
DOI: 10.1016/j.automatica.2012.06.051
ISSN: 0005-1098
eISSN: 1873-2836
Grant note: DE-FG52-09NA29364 / DoE
Language: English
Date published: 09/2012
Academic Unit: Statistics and Actuarial Science; Civil and Environmental Engineering; Electrical and Computer Engineering; Radiology
Record Identifier: 9983985818502771

Metrics

18 Record Views

20 Times Cited - Web of Science