Selecting negative examples for hierarchical text classification: An experimental comparison

Tiziano Fagni, Fabrizio Sebastiani

Research output: Contribution to journalArticle

7 Citations (Scopus)

Abstract

Hierarchical text classification (HTC) approaches have recently attracted a lot of interest on the part of researchers in human language technology and machine learning, since they have been shown to bring about equal, if not better, classification accuracy with respect to their "flat" counterparts while allowing exponential time savings at both learning and classification time. A typical component of HTC methods is a "local" policy for selecting negative examples: Given a category c, its negative training examples are by default identified with the training examples that are negative for c and positive for the categories which are siblings of c in the hierarchy. However, this policy has always been taken for granted and never been subjected to careful scrutiny since first proposed 15 years ago. This article proposes a thorough experimental comparison between this policy and three other policies for the selection of negative examples in HTC contexts, one of which (BEST LOCAL (k)) is being proposed for the first time in this article. We compare these policies on the hierarchical versions of three supervised learning algorithms (boosting, support vector machines, and naïve Bayes) by performing experiments on two standard TC datasets, REUTERS-21578 and RCV1-V2.

Original languageEnglish
Pages (from-to)2256-2265
Number of pages10
JournalJournal of the American Society for Information Science and Technology
Volume61
Issue number11
DOIs
Publication statusPublished - Nov 2010
Externally publishedYes

Fingerprint

learning
Supervised learning
Learning algorithms
Support vector machines
savings
Learning systems
Text classification
experiment
language
Experiments
time
Boosting
Savings
Learning algorithm
Machine learning
Support vector machine
Siblings
Language
Experiment

ASJC Scopus subject areas

  • Software
  • Artificial Intelligence
  • Information Systems
  • Human-Computer Interaction
  • Computer Networks and Communications

Cite this

Selecting negative examples for hierarchical text classification : An experimental comparison. / Fagni, Tiziano; Sebastiani, Fabrizio.

In: Journal of the American Society for Information Science and Technology, Vol. 61, No. 11, 11.2010, p. 2256-2265.

Research output: Contribution to journalArticle

@article{6504b3ecc71546888a3af9952cade16f,
title = "Selecting negative examples for hierarchical text classification: An experimental comparison",
abstract = "Hierarchical text classification (HTC) approaches have recently attracted a lot of interest on the part of researchers in human language technology and machine learning, since they have been shown to bring about equal, if not better, classification accuracy with respect to their {"}flat{"} counterparts while allowing exponential time savings at both learning and classification time. A typical component of HTC methods is a {"}local{"} policy for selecting negative examples: Given a category c, its negative training examples are by default identified with the training examples that are negative for c and positive for the categories which are siblings of c in the hierarchy. However, this policy has always been taken for granted and never been subjected to careful scrutiny since first proposed 15 years ago. This article proposes a thorough experimental comparison between this policy and three other policies for the selection of negative examples in HTC contexts, one of which (BEST LOCAL (k)) is being proposed for the first time in this article. We compare these policies on the hierarchical versions of three supervised learning algorithms (boosting, support vector machines, and na{\"i}ve Bayes) by performing experiments on two standard TC datasets, REUTERS-21578 and RCV1-V2.",
author = "Tiziano Fagni and Fabrizio Sebastiani",
year = "2010",
month = "11",
doi = "10.1002/asi.21411",
language = "English",
volume = "61",
pages = "2256--2265",
journal = "Journal of the Association for Information Science and Technology",
issn = "2330-1635",
publisher = "John Wiley and Sons Ltd",
number = "11",

}

TY - JOUR

T1 - Selecting negative examples for hierarchical text classification

T2 - An experimental comparison

AU - Fagni, Tiziano

AU - Sebastiani, Fabrizio

PY - 2010/11

Y1 - 2010/11

N2 - Hierarchical text classification (HTC) approaches have recently attracted a lot of interest on the part of researchers in human language technology and machine learning, since they have been shown to bring about equal, if not better, classification accuracy with respect to their "flat" counterparts while allowing exponential time savings at both learning and classification time. A typical component of HTC methods is a "local" policy for selecting negative examples: Given a category c, its negative training examples are by default identified with the training examples that are negative for c and positive for the categories which are siblings of c in the hierarchy. However, this policy has always been taken for granted and never been subjected to careful scrutiny since first proposed 15 years ago. This article proposes a thorough experimental comparison between this policy and three other policies for the selection of negative examples in HTC contexts, one of which (BEST LOCAL (k)) is being proposed for the first time in this article. We compare these policies on the hierarchical versions of three supervised learning algorithms (boosting, support vector machines, and naïve Bayes) by performing experiments on two standard TC datasets, REUTERS-21578 and RCV1-V2.

AB - Hierarchical text classification (HTC) approaches have recently attracted a lot of interest on the part of researchers in human language technology and machine learning, since they have been shown to bring about equal, if not better, classification accuracy with respect to their "flat" counterparts while allowing exponential time savings at both learning and classification time. A typical component of HTC methods is a "local" policy for selecting negative examples: Given a category c, its negative training examples are by default identified with the training examples that are negative for c and positive for the categories which are siblings of c in the hierarchy. However, this policy has always been taken for granted and never been subjected to careful scrutiny since first proposed 15 years ago. This article proposes a thorough experimental comparison between this policy and three other policies for the selection of negative examples in HTC contexts, one of which (BEST LOCAL (k)) is being proposed for the first time in this article. We compare these policies on the hierarchical versions of three supervised learning algorithms (boosting, support vector machines, and naïve Bayes) by performing experiments on two standard TC datasets, REUTERS-21578 and RCV1-V2.

UR - http://www.scopus.com/inward/record.url?scp=78149454350&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=78149454350&partnerID=8YFLogxK

U2 - 10.1002/asi.21411

DO - 10.1002/asi.21411

M3 - Article

AN - SCOPUS:78149454350

VL - 61

SP - 2256

EP - 2265

JO - Journal of the Association for Information Science and Technology

JF - Journal of the Association for Information Science and Technology

SN - 2330-1635

IS - 11

ER -