An Evidential Approach to Classification Combination for Text Categorisation

D. Bell, J. Guan, Yaxin Bi

Research output: Contribution to journalArticle

Abstract

In this paper we look at a way of combining two or more different classification methods for text categorization. The specific methods we have been experimenting with in our group include the Support Vector Machine, kNN (nearest neighbours), kNN model-based approach (kNNM), and Rocchio methods. Then we describe our method for combining the classifiers. A previous study suggested that the combination of the best and the second best classifiers using evidential operations [1] can achieve better performance than other combinations. We assess some aspects of this from an evidential reasoning perspective and suggest a refinement of the approach.
LanguageEnglish
Pages13-22
JournalKnowledge Mining
Volume185
DOIs
Publication statusPublished - 2005

Fingerprint

Classifiers
Support vector machines

Cite this

@article{caefa264282b4ebda2d61d25b754b8c5,
title = "An Evidential Approach to Classification Combination for Text Categorisation",
abstract = "In this paper we look at a way of combining two or more different classification methods for text categorization. The specific methods we have been experimenting with in our group include the Support Vector Machine, kNN (nearest neighbours), kNN model-based approach (kNNM), and Rocchio methods. Then we describe our method for combining the classifiers. A previous study suggested that the combination of the best and the second best classifiers using evidential operations [1] can achieve better performance than other combinations. We assess some aspects of this from an evidential reasoning perspective and suggest a refinement of the approach.",
author = "D. Bell and J. Guan and Yaxin Bi",
year = "2005",
doi = "10.1007/3-540-32394-5_2",
language = "English",
volume = "185",
pages = "13--22",

}

An Evidential Approach to Classification Combination for Text Categorisation. / Bell, D.; Guan, J.; Bi, Yaxin.

Vol. 185, 2005, p. 13-22.

Research output: Contribution to journalArticle

TY - JOUR

T1 - An Evidential Approach to Classification Combination for Text Categorisation

AU - Bell, D.

AU - Guan, J.

AU - Bi, Yaxin

PY - 2005

Y1 - 2005

N2 - In this paper we look at a way of combining two or more different classification methods for text categorization. The specific methods we have been experimenting with in our group include the Support Vector Machine, kNN (nearest neighbours), kNN model-based approach (kNNM), and Rocchio methods. Then we describe our method for combining the classifiers. A previous study suggested that the combination of the best and the second best classifiers using evidential operations [1] can achieve better performance than other combinations. We assess some aspects of this from an evidential reasoning perspective and suggest a refinement of the approach.

AB - In this paper we look at a way of combining two or more different classification methods for text categorization. The specific methods we have been experimenting with in our group include the Support Vector Machine, kNN (nearest neighbours), kNN model-based approach (kNNM), and Rocchio methods. Then we describe our method for combining the classifiers. A previous study suggested that the combination of the best and the second best classifiers using evidential operations [1] can achieve better performance than other combinations. We assess some aspects of this from an evidential reasoning perspective and suggest a refinement of the approach.

U2 - 10.1007/3-540-32394-5_2

DO - 10.1007/3-540-32394-5_2

M3 - Article

VL - 185

SP - 13

EP - 22

ER -