A comparison of the classical and the linear programming approaches to the classification problem in discriminant analysis

Ravinder Nath, Wade M. Jackson, Thomas W. Jones

Research output: Contribution to journalArticle

20 Citations (Scopus)

Abstract

Several mathematical programming approaches to the classification problem in discriminant analysis have recently been introduced. This paper empirically compares these newly introduced classification techniques with Fisher's linear discriminant analysis (FLDA). quadratic discriminant analysis (QDA), logit analysis, and several rank-based procedures for a variety of symmetric and skewed distributions. The percent of correctly classified observations by each procedure in a holdout sample indicate that while under some experimental conditions the linear programming approaches compete well with the classical procedures, overall, however, their performance lags behind that of the classical procedures.

Original languageEnglish
Pages (from-to)73-93
Number of pages21
JournalJournal of Statistical Computation and Simulation
Volume41
Issue number1-2
DOIs
StatePublished - May 1 1992
Externally publishedYes

Fingerprint

Discriminant analysis
Discriminant Analysis
Classification Problems
Linear programming
Mathematical programming
Skewed Distribution
Logit
Symmetric Distributions
Mathematical Programming
Percent

All Science Journal Classification (ASJC) codes

  • Statistics, Probability and Uncertainty
  • Modeling and Simulation
  • Statistics and Probability
  • Applied Mathematics

Cite this

A comparison of the classical and the linear programming approaches to the classification problem in discriminant analysis. / Nath, Ravinder; Jackson, Wade M.; Jones, Thomas W.

In: Journal of Statistical Computation and Simulation, Vol. 41, No. 1-2, 01.05.1992, p. 73-93.

Research output: Contribution to journalArticle

@article{0a2ac4ec2c384fa69a13511060d622a6,
title = "A comparison of the classical and the linear programming approaches to the classification problem in discriminant analysis",
abstract = "Several mathematical programming approaches to the classification problem in discriminant analysis have recently been introduced. This paper empirically compares these newly introduced classification techniques with Fisher's linear discriminant analysis (FLDA). quadratic discriminant analysis (QDA), logit analysis, and several rank-based procedures for a variety of symmetric and skewed distributions. The percent of correctly classified observations by each procedure in a holdout sample indicate that while under some experimental conditions the linear programming approaches compete well with the classical procedures, overall, however, their performance lags behind that of the classical procedures.",
author = "Ravinder Nath and Jackson, {Wade M.} and Jones, {Thomas W.}",
year = "1992",
month = "5",
day = "1",
doi = "10.1080/00949659208811392",
language = "English",
volume = "41",
pages = "73--93",
journal = "Journal of Statistical Computation and Simulation",
issn = "0094-9655",
publisher = "Taylor and Francis Ltd.",
number = "1-2",

}

TY - JOUR

T1 - A comparison of the classical and the linear programming approaches to the classification problem in discriminant analysis

AU - Nath, Ravinder

AU - Jackson, Wade M.

AU - Jones, Thomas W.

PY - 1992/5/1

Y1 - 1992/5/1

N2 - Several mathematical programming approaches to the classification problem in discriminant analysis have recently been introduced. This paper empirically compares these newly introduced classification techniques with Fisher's linear discriminant analysis (FLDA). quadratic discriminant analysis (QDA), logit analysis, and several rank-based procedures for a variety of symmetric and skewed distributions. The percent of correctly classified observations by each procedure in a holdout sample indicate that while under some experimental conditions the linear programming approaches compete well with the classical procedures, overall, however, their performance lags behind that of the classical procedures.

AB - Several mathematical programming approaches to the classification problem in discriminant analysis have recently been introduced. This paper empirically compares these newly introduced classification techniques with Fisher's linear discriminant analysis (FLDA). quadratic discriminant analysis (QDA), logit analysis, and several rank-based procedures for a variety of symmetric and skewed distributions. The percent of correctly classified observations by each procedure in a holdout sample indicate that while under some experimental conditions the linear programming approaches compete well with the classical procedures, overall, however, their performance lags behind that of the classical procedures.

UR - http://www.scopus.com/inward/record.url?scp=0011850273&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=0011850273&partnerID=8YFLogxK

U2 - 10.1080/00949659208811392

DO - 10.1080/00949659208811392

M3 - Article

VL - 41

SP - 73

EP - 93

JO - Journal of Statistical Computation and Simulation

JF - Journal of Statistical Computation and Simulation

SN - 0094-9655

IS - 1-2

ER -