Choosing the best set of variables in regression analysis using integer programming

Hiroshi Konno, Rei Yamamoto

Research output: Contribution to journalArticle

17 Citations (Scopus)

Abstract

This paper is concerned with an algorithm for selecting the best set of s variables out of k(> s) candidate variables in a multiple linear regression model. We employ absolute deviation as the measure of deviation and solve the resulting optimization problem by using 0-1 integer programming methodologies. In addition, we will propose a heuristic algorithm to obtain a close to optimal set of variables in terms of squared deviation. Computational results show that this method is practical and reliable for determining the best set of variables.

Original languageEnglish
Pages (from-to)273-282
Number of pages10
JournalJournal of Global Optimization
Volume44
Issue number2
DOIs
Publication statusPublished - 2009 Jun 1
Externally publishedYes

Fingerprint

Integer programming
Heuristic algorithms
Integer Programming
Regression Analysis
Linear regression
Regression analysis
Deviation
0-1 Integer Programming
Multiple Linear Regression
Linear Regression Model
Heuristic algorithm
Computational Results
Optimization Problem
Methodology

Keywords

  • 0-1 integer programming
  • Cardinality constraint
  • Least absolute deviation
  • Linear regression
  • Variable selection

ASJC Scopus subject areas

  • Computer Science Applications
  • Management Science and Operations Research
  • Applied Mathematics
  • Control and Optimization

Cite this

Choosing the best set of variables in regression analysis using integer programming. / Konno, Hiroshi; Yamamoto, Rei.

In: Journal of Global Optimization, Vol. 44, No. 2, 01.06.2009, p. 273-282.

Research output: Contribution to journalArticle

@article{38367c24c84244318bf85d6db4b53e98,
title = "Choosing the best set of variables in regression analysis using integer programming",
abstract = "This paper is concerned with an algorithm for selecting the best set of s variables out of k(> s) candidate variables in a multiple linear regression model. We employ absolute deviation as the measure of deviation and solve the resulting optimization problem by using 0-1 integer programming methodologies. In addition, we will propose a heuristic algorithm to obtain a close to optimal set of variables in terms of squared deviation. Computational results show that this method is practical and reliable for determining the best set of variables.",
keywords = "0-1 integer programming, Cardinality constraint, Least absolute deviation, Linear regression, Variable selection",
author = "Hiroshi Konno and Rei Yamamoto",
year = "2009",
month = "6",
day = "1",
doi = "10.1007/s10898-008-9323-9",
language = "English",
volume = "44",
pages = "273--282",
journal = "Journal of Global Optimization",
issn = "0925-5001",
publisher = "Springer Netherlands",
number = "2",

}

TY - JOUR

T1 - Choosing the best set of variables in regression analysis using integer programming

AU - Konno, Hiroshi

AU - Yamamoto, Rei

PY - 2009/6/1

Y1 - 2009/6/1

N2 - This paper is concerned with an algorithm for selecting the best set of s variables out of k(> s) candidate variables in a multiple linear regression model. We employ absolute deviation as the measure of deviation and solve the resulting optimization problem by using 0-1 integer programming methodologies. In addition, we will propose a heuristic algorithm to obtain a close to optimal set of variables in terms of squared deviation. Computational results show that this method is practical and reliable for determining the best set of variables.

AB - This paper is concerned with an algorithm for selecting the best set of s variables out of k(> s) candidate variables in a multiple linear regression model. We employ absolute deviation as the measure of deviation and solve the resulting optimization problem by using 0-1 integer programming methodologies. In addition, we will propose a heuristic algorithm to obtain a close to optimal set of variables in terms of squared deviation. Computational results show that this method is practical and reliable for determining the best set of variables.

KW - 0-1 integer programming

KW - Cardinality constraint

KW - Least absolute deviation

KW - Linear regression

KW - Variable selection

UR - http://www.scopus.com/inward/record.url?scp=67349161084&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=67349161084&partnerID=8YFLogxK

U2 - 10.1007/s10898-008-9323-9

DO - 10.1007/s10898-008-9323-9

M3 - Article

AN - SCOPUS:67349161084

VL - 44

SP - 273

EP - 282

JO - Journal of Global Optimization

JF - Journal of Global Optimization

SN - 0925-5001

IS - 2

ER -