Effective use of WordNet semantics via kernel-based learning

Roberto Basili, Marco Cammisa, Alessandro Moschitti

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

24 Citations (Scopus)

Abstract

Research on document similarity has shown that complex representations are not more accurate than the simple bag-ofwords. Term clustering, e.g. using latent semantic indexing, word co-occurrences or synonym relations using a word ontology have been shown not very effective. In particular, when to extend the similarity function external prior knowledge is used, e.g. WordNet, the retrieval system decreases its performance. The critical issues here are methods and conditions to integrate such knowledge. In this paper we propose kernel functions to add prior knowledge to learning algorithms for document classification. Such kernels use a term similarity measure based on the WordNet hierarchy. The kernel trick is used to implement such space in a balanced and statistically coherent way. Cross-validation results show the benefit of the approach for the Support Vector Machines when few training data is available.

Original languageEnglish
Title of host publicationCoNLL 2005 - Proceedings of the Ninth Conference on Computational Natural Language Learning
Pages1-8
Number of pages8
Publication statusPublished - 1 Dec 2005
Externally publishedYes
Event9th Conference on Computational Natural Language Learning, CoNLL 2005 - Ann Arbor, MI, United States
Duration: 29 Jun 200530 Jun 2005

Other

Other9th Conference on Computational Natural Language Learning, CoNLL 2005
CountryUnited States
CityAnn Arbor, MI
Period29/6/0530/6/05

Fingerprint

Semantics
semantics
indexing
ontology
Learning algorithms
knowledge
learning
Support vector machines
Ontology
performance

ASJC Scopus subject areas

  • Artificial Intelligence
  • Human-Computer Interaction
  • Linguistics and Language

Cite this

Basili, R., Cammisa, M., & Moschitti, A. (2005). Effective use of WordNet semantics via kernel-based learning. In CoNLL 2005 - Proceedings of the Ninth Conference on Computational Natural Language Learning (pp. 1-8)

Effective use of WordNet semantics via kernel-based learning. / Basili, Roberto; Cammisa, Marco; Moschitti, Alessandro.

CoNLL 2005 - Proceedings of the Ninth Conference on Computational Natural Language Learning. 2005. p. 1-8.

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

Basili, R, Cammisa, M & Moschitti, A 2005, Effective use of WordNet semantics via kernel-based learning. in CoNLL 2005 - Proceedings of the Ninth Conference on Computational Natural Language Learning. pp. 1-8, 9th Conference on Computational Natural Language Learning, CoNLL 2005, Ann Arbor, MI, United States, 29/6/05.
Basili R, Cammisa M, Moschitti A. Effective use of WordNet semantics via kernel-based learning. In CoNLL 2005 - Proceedings of the Ninth Conference on Computational Natural Language Learning. 2005. p. 1-8
Basili, Roberto ; Cammisa, Marco ; Moschitti, Alessandro. / Effective use of WordNet semantics via kernel-based learning. CoNLL 2005 - Proceedings of the Ninth Conference on Computational Natural Language Learning. 2005. pp. 1-8
@inproceedings{6ae253bff1114d12b8aad0e2d722f121,
title = "Effective use of WordNet semantics via kernel-based learning",
abstract = "Research on document similarity has shown that complex representations are not more accurate than the simple bag-ofwords. Term clustering, e.g. using latent semantic indexing, word co-occurrences or synonym relations using a word ontology have been shown not very effective. In particular, when to extend the similarity function external prior knowledge is used, e.g. WordNet, the retrieval system decreases its performance. The critical issues here are methods and conditions to integrate such knowledge. In this paper we propose kernel functions to add prior knowledge to learning algorithms for document classification. Such kernels use a term similarity measure based on the WordNet hierarchy. The kernel trick is used to implement such space in a balanced and statistically coherent way. Cross-validation results show the benefit of the approach for the Support Vector Machines when few training data is available.",
author = "Roberto Basili and Marco Cammisa and Alessandro Moschitti",
year = "2005",
month = "12",
day = "1",
language = "English",
pages = "1--8",
booktitle = "CoNLL 2005 - Proceedings of the Ninth Conference on Computational Natural Language Learning",

}

TY - GEN

T1 - Effective use of WordNet semantics via kernel-based learning

AU - Basili, Roberto

AU - Cammisa, Marco

AU - Moschitti, Alessandro

PY - 2005/12/1

Y1 - 2005/12/1

N2 - Research on document similarity has shown that complex representations are not more accurate than the simple bag-ofwords. Term clustering, e.g. using latent semantic indexing, word co-occurrences or synonym relations using a word ontology have been shown not very effective. In particular, when to extend the similarity function external prior knowledge is used, e.g. WordNet, the retrieval system decreases its performance. The critical issues here are methods and conditions to integrate such knowledge. In this paper we propose kernel functions to add prior knowledge to learning algorithms for document classification. Such kernels use a term similarity measure based on the WordNet hierarchy. The kernel trick is used to implement such space in a balanced and statistically coherent way. Cross-validation results show the benefit of the approach for the Support Vector Machines when few training data is available.

AB - Research on document similarity has shown that complex representations are not more accurate than the simple bag-ofwords. Term clustering, e.g. using latent semantic indexing, word co-occurrences or synonym relations using a word ontology have been shown not very effective. In particular, when to extend the similarity function external prior knowledge is used, e.g. WordNet, the retrieval system decreases its performance. The critical issues here are methods and conditions to integrate such knowledge. In this paper we propose kernel functions to add prior knowledge to learning algorithms for document classification. Such kernels use a term similarity measure based on the WordNet hierarchy. The kernel trick is used to implement such space in a balanced and statistically coherent way. Cross-validation results show the benefit of the approach for the Support Vector Machines when few training data is available.

UR - http://www.scopus.com/inward/record.url?scp=84862298382&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84862298382&partnerID=8YFLogxK

M3 - Conference contribution

AN - SCOPUS:84862298382

SP - 1

EP - 8

BT - CoNLL 2005 - Proceedings of the Ninth Conference on Computational Natural Language Learning

ER -