Learning to rank from structures in hierarchical text classification

Qi Ju, Alessandro Moschitti, Richard Johansson

Research output: Chapter in Book/Report/Conference proceedingConference contribution

2 Citations (Scopus)

Abstract

In this paper, we model learning to rank algorithms based on structural dependencies in hierarchical multi-label text categorization (TC). Our method uses the classification probability of the binary classifiers of a standard top-down approach to generate k-best hypotheses. The latter are generated according to their global probability while at the same time satisfy the structural constraints between father and children nodes. The rank is then refined using Support Vector Machines and tree kernels applied to a structural representation of hypotheses, i.e., a hierarchy tree in which the outcome of binary one-vs-all classifiers is directly marked in its nodes. Our extensive experiments on the whole Reuters Corpus Volume 1 show that our models significantly improve over the state of the art in TC, thanks to the use of structural dependecies.

Original languageEnglish
Title of host publicationLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Pages183-194
Number of pages12
Volume7814 LNCS
DOIs
Publication statusPublished - 2 Apr 2013
Externally publishedYes
Event35th European Conference on Information Retrieval, ECIR 2013 - Moscow, Russian Federation
Duration: 24 Mar 201327 Mar 2013

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume7814 LNCS
ISSN (Print)03029743
ISSN (Electronic)16113349

Other

Other35th European Conference on Information Retrieval, ECIR 2013
CountryRussian Federation
CityMoscow
Period24/3/1327/3/13

Fingerprint

Hierarchical Classification
Text Categorization
Text Classification
Classifiers
Classifier
Binary
Vertex of a graph
Support vector machines
Labels
Support Vector Machine
kernel
Model
Experiment
Experiments
Learning
Hierarchy
Children
Corpus
Standards

ASJC Scopus subject areas

  • Computer Science(all)
  • Theoretical Computer Science

Cite this

Ju, Q., Moschitti, A., & Johansson, R. (2013). Learning to rank from structures in hierarchical text classification. In Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) (Vol. 7814 LNCS, pp. 183-194). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 7814 LNCS). https://doi.org/10.1007/978-3-642-36973-5_16

Learning to rank from structures in hierarchical text classification. / Ju, Qi; Moschitti, Alessandro; Johansson, Richard.

Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics). Vol. 7814 LNCS 2013. p. 183-194 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 7814 LNCS).

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Ju, Q, Moschitti, A & Johansson, R 2013, Learning to rank from structures in hierarchical text classification. in Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics). vol. 7814 LNCS, Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), vol. 7814 LNCS, pp. 183-194, 35th European Conference on Information Retrieval, ECIR 2013, Moscow, Russian Federation, 24/3/13. https://doi.org/10.1007/978-3-642-36973-5_16
Ju Q, Moschitti A, Johansson R. Learning to rank from structures in hierarchical text classification. In Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics). Vol. 7814 LNCS. 2013. p. 183-194. (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)). https://doi.org/10.1007/978-3-642-36973-5_16
Ju, Qi ; Moschitti, Alessandro ; Johansson, Richard. / Learning to rank from structures in hierarchical text classification. Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics). Vol. 7814 LNCS 2013. pp. 183-194 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)).
@inproceedings{d3e581591d92482e93b8de461d87b217,
title = "Learning to rank from structures in hierarchical text classification",
abstract = "In this paper, we model learning to rank algorithms based on structural dependencies in hierarchical multi-label text categorization (TC). Our method uses the classification probability of the binary classifiers of a standard top-down approach to generate k-best hypotheses. The latter are generated according to their global probability while at the same time satisfy the structural constraints between father and children nodes. The rank is then refined using Support Vector Machines and tree kernels applied to a structural representation of hypotheses, i.e., a hierarchy tree in which the outcome of binary one-vs-all classifiers is directly marked in its nodes. Our extensive experiments on the whole Reuters Corpus Volume 1 show that our models significantly improve over the state of the art in TC, thanks to the use of structural dependecies.",
author = "Qi Ju and Alessandro Moschitti and Richard Johansson",
year = "2013",
month = "4",
day = "2",
doi = "10.1007/978-3-642-36973-5_16",
language = "English",
isbn = "9783642369728",
volume = "7814 LNCS",
series = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",
pages = "183--194",
booktitle = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",

}

TY - GEN

T1 - Learning to rank from structures in hierarchical text classification

AU - Ju, Qi

AU - Moschitti, Alessandro

AU - Johansson, Richard

PY - 2013/4/2

Y1 - 2013/4/2

N2 - In this paper, we model learning to rank algorithms based on structural dependencies in hierarchical multi-label text categorization (TC). Our method uses the classification probability of the binary classifiers of a standard top-down approach to generate k-best hypotheses. The latter are generated according to their global probability while at the same time satisfy the structural constraints between father and children nodes. The rank is then refined using Support Vector Machines and tree kernels applied to a structural representation of hypotheses, i.e., a hierarchy tree in which the outcome of binary one-vs-all classifiers is directly marked in its nodes. Our extensive experiments on the whole Reuters Corpus Volume 1 show that our models significantly improve over the state of the art in TC, thanks to the use of structural dependecies.

AB - In this paper, we model learning to rank algorithms based on structural dependencies in hierarchical multi-label text categorization (TC). Our method uses the classification probability of the binary classifiers of a standard top-down approach to generate k-best hypotheses. The latter are generated according to their global probability while at the same time satisfy the structural constraints between father and children nodes. The rank is then refined using Support Vector Machines and tree kernels applied to a structural representation of hypotheses, i.e., a hierarchy tree in which the outcome of binary one-vs-all classifiers is directly marked in its nodes. Our extensive experiments on the whole Reuters Corpus Volume 1 show that our models significantly improve over the state of the art in TC, thanks to the use of structural dependecies.

UR - http://www.scopus.com/inward/record.url?scp=84875427676&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84875427676&partnerID=8YFLogxK

U2 - 10.1007/978-3-642-36973-5_16

DO - 10.1007/978-3-642-36973-5_16

M3 - Conference contribution

SN - 9783642369728

VL - 7814 LNCS

T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)

SP - 183

EP - 194

BT - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)

ER -