A kernel fisher discriminant analysis-based tree ensemble classifier: KFDA forest

Donghwan Kim, Seung Hwan Park, Jun-Geol Baek

Research output: Contribution to journalArticle

Abstract

In general, an ensemble classifier is more accurate than a single classifier. In this study, we propose an ensemble classifier called the kernel Fisher discriminant analysis forest (KFDA Forest), which is a tree-based ensemble method that applies KFDA. To promote diversity, bootstrap is used, and variable sets are randomly divided into K subsets. KFDA is performed on each subset to increase classification accuracy. KFDA maximizes the distance between classes while minimizing the distance within classes. KFDA can also be applied to classification problems in a nonlinear data structure using the kernel trick because it can transform the input space into a kernel feature space, commonly named a rotation, rather than performing a dimensionality reduction. Because new feature axes and KFDA projections are parallel, decision trees are used as a base classifier. To compare the proposed method with existing ensemble methods, we apply these to real datasets from the UCI and KEEL repositories.

Original languageEnglish
Pages (from-to)569-579
Number of pages11
JournalInternational Journal of Industrial Engineering : Theory Applications and Practice
Volume25
Publication statusPublished - 2018 Jan 1

Fingerprint

Discriminant analysis
Classifiers
Decision trees
Data structures

Keywords

  • Classification
  • Decision trees
  • Ensemble classifier
  • Kernel fisher discriminant analysis
  • Rotation forest

ASJC Scopus subject areas

  • Industrial and Manufacturing Engineering

Cite this

A kernel fisher discriminant analysis-based tree ensemble classifier : KFDA forest. / Kim, Donghwan; Park, Seung Hwan; Baek, Jun-Geol.

In: International Journal of Industrial Engineering : Theory Applications and Practice, Vol. 25, 01.01.2018, p. 569-579.

Research output: Contribution to journalArticle

@article{5ab88b01edc24a3880f26aecafb189a6,
title = "A kernel fisher discriminant analysis-based tree ensemble classifier: KFDA forest",
abstract = "In general, an ensemble classifier is more accurate than a single classifier. In this study, we propose an ensemble classifier called the kernel Fisher discriminant analysis forest (KFDA Forest), which is a tree-based ensemble method that applies KFDA. To promote diversity, bootstrap is used, and variable sets are randomly divided into K subsets. KFDA is performed on each subset to increase classification accuracy. KFDA maximizes the distance between classes while minimizing the distance within classes. KFDA can also be applied to classification problems in a nonlinear data structure using the kernel trick because it can transform the input space into a kernel feature space, commonly named a rotation, rather than performing a dimensionality reduction. Because new feature axes and KFDA projections are parallel, decision trees are used as a base classifier. To compare the proposed method with existing ensemble methods, we apply these to real datasets from the UCI and KEEL repositories.",
keywords = "Classification, Decision trees, Ensemble classifier, Kernel fisher discriminant analysis, Rotation forest",
author = "Donghwan Kim and Park, {Seung Hwan} and Jun-Geol Baek",
year = "2018",
month = "1",
day = "1",
language = "English",
volume = "25",
pages = "569--579",
journal = "International Journal of Industrial Engineering : Theory Applications and Practice",
issn = "1072-4761",
publisher = "University of Cincinnati",

}

TY - JOUR

T1 - A kernel fisher discriminant analysis-based tree ensemble classifier

T2 - KFDA forest

AU - Kim, Donghwan

AU - Park, Seung Hwan

AU - Baek, Jun-Geol

PY - 2018/1/1

Y1 - 2018/1/1

N2 - In general, an ensemble classifier is more accurate than a single classifier. In this study, we propose an ensemble classifier called the kernel Fisher discriminant analysis forest (KFDA Forest), which is a tree-based ensemble method that applies KFDA. To promote diversity, bootstrap is used, and variable sets are randomly divided into K subsets. KFDA is performed on each subset to increase classification accuracy. KFDA maximizes the distance between classes while minimizing the distance within classes. KFDA can also be applied to classification problems in a nonlinear data structure using the kernel trick because it can transform the input space into a kernel feature space, commonly named a rotation, rather than performing a dimensionality reduction. Because new feature axes and KFDA projections are parallel, decision trees are used as a base classifier. To compare the proposed method with existing ensemble methods, we apply these to real datasets from the UCI and KEEL repositories.

AB - In general, an ensemble classifier is more accurate than a single classifier. In this study, we propose an ensemble classifier called the kernel Fisher discriminant analysis forest (KFDA Forest), which is a tree-based ensemble method that applies KFDA. To promote diversity, bootstrap is used, and variable sets are randomly divided into K subsets. KFDA is performed on each subset to increase classification accuracy. KFDA maximizes the distance between classes while minimizing the distance within classes. KFDA can also be applied to classification problems in a nonlinear data structure using the kernel trick because it can transform the input space into a kernel feature space, commonly named a rotation, rather than performing a dimensionality reduction. Because new feature axes and KFDA projections are parallel, decision trees are used as a base classifier. To compare the proposed method with existing ensemble methods, we apply these to real datasets from the UCI and KEEL repositories.

KW - Classification

KW - Decision trees

KW - Ensemble classifier

KW - Kernel fisher discriminant analysis

KW - Rotation forest

UR - http://www.scopus.com/inward/record.url?scp=85060724625&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85060724625&partnerID=8YFLogxK

M3 - Article

AN - SCOPUS:85060724625

VL - 25

SP - 569

EP - 579

JO - International Journal of Industrial Engineering : Theory Applications and Practice

JF - International Journal of Industrial Engineering : Theory Applications and Practice

SN - 1072-4761

ER -