RV-SVM: An efficient method for learning ranking SVM

Yu Hwanjo, Kim Youngdae, Hwang Seungwon

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

8 Citations (Scopus)

Abstract

Learning ranking (or preference) functions has become an important data mining task in recent years, as various applications have been found in information retrieval. Among rank learning methods, ranking SVMhas been favorably applied to various applications, e.g., optimizing search engines, improving data retrieval quality. In this paper, we first develop a 1-norm ranking SVM that is faster in testing than the standard ranking SVM, and propose Ranking Vector SVM (RV-SVM) that revises the 1-norm ranking SVM for faster training. The number of variables in the RV-SVM is significantly smaller, thus the RV-SVM trains much faster than the other ranking SVMs.We experimentally compared the RV-SVM with the state-of-the-art rank learning method provided in SVM-light. The RV-SVMuses much less support vectors and trains much faster for nonlinear kernels than the SVM-light. The accuracies of RV-SVM and SVM-light are comparable on relatively large data sets. Our implementation of RV-SVM is posted at http://iis.postech.ac.kr/rv-svm.

Original languageEnglish
Title of host publication13th Pacific-Asia Conference on Knowledge Discovery and Data Mining, PAKDD 2009
Pages426-438
Number of pages13
DOIs
Publication statusPublished - 2009 Jul 23
Event13th Pacific-Asia Conference on Knowledge Discovery and Data Mining, PAKDD 2009 - Bangkok, Thailand
Duration: 2009 Apr 272009 Apr 30

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume5476 LNAI
ISSN (Print)0302-9743
ISSN (Electronic)1611-3349

Other

Other13th Pacific-Asia Conference on Knowledge Discovery and Data Mining, PAKDD 2009
CountryThailand
CityBangkok
Period09/4/2709/4/30

Fingerprint

Ranking
Search engines
Learning
Information retrieval
Data mining
Norm
Support Vector
Search Engine
Large Data Sets
Information Retrieval
Testing
Data Mining
Retrieval
kernel

All Science Journal Classification (ASJC) codes

  • Theoretical Computer Science
  • Computer Science(all)

Cite this

Hwanjo, Y., Youngdae, K., & Seungwon, H. (2009). RV-SVM: An efficient method for learning ranking SVM. In 13th Pacific-Asia Conference on Knowledge Discovery and Data Mining, PAKDD 2009 (pp. 426-438). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 5476 LNAI). https://doi.org/10.1007/978-3-642-01307-2_39
Hwanjo, Yu ; Youngdae, Kim ; Seungwon, Hwang. / RV-SVM : An efficient method for learning ranking SVM. 13th Pacific-Asia Conference on Knowledge Discovery and Data Mining, PAKDD 2009. 2009. pp. 426-438 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)).
@inproceedings{df1b25b07ff74d018779cd34c1cd1e93,
title = "RV-SVM: An efficient method for learning ranking SVM",
abstract = "Learning ranking (or preference) functions has become an important data mining task in recent years, as various applications have been found in information retrieval. Among rank learning methods, ranking SVMhas been favorably applied to various applications, e.g., optimizing search engines, improving data retrieval quality. In this paper, we first develop a 1-norm ranking SVM that is faster in testing than the standard ranking SVM, and propose Ranking Vector SVM (RV-SVM) that revises the 1-norm ranking SVM for faster training. The number of variables in the RV-SVM is significantly smaller, thus the RV-SVM trains much faster than the other ranking SVMs.We experimentally compared the RV-SVM with the state-of-the-art rank learning method provided in SVM-light. The RV-SVMuses much less support vectors and trains much faster for nonlinear kernels than the SVM-light. The accuracies of RV-SVM and SVM-light are comparable on relatively large data sets. Our implementation of RV-SVM is posted at http://iis.postech.ac.kr/rv-svm.",
author = "Yu Hwanjo and Kim Youngdae and Hwang Seungwon",
year = "2009",
month = "7",
day = "23",
doi = "10.1007/978-3-642-01307-2_39",
language = "English",
isbn = "3642013066",
series = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",
pages = "426--438",
booktitle = "13th Pacific-Asia Conference on Knowledge Discovery and Data Mining, PAKDD 2009",

}

Hwanjo, Y, Youngdae, K & Seungwon, H 2009, RV-SVM: An efficient method for learning ranking SVM. in 13th Pacific-Asia Conference on Knowledge Discovery and Data Mining, PAKDD 2009. Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), vol. 5476 LNAI, pp. 426-438, 13th Pacific-Asia Conference on Knowledge Discovery and Data Mining, PAKDD 2009, Bangkok, Thailand, 09/4/27. https://doi.org/10.1007/978-3-642-01307-2_39

RV-SVM : An efficient method for learning ranking SVM. / Hwanjo, Yu; Youngdae, Kim; Seungwon, Hwang.

13th Pacific-Asia Conference on Knowledge Discovery and Data Mining, PAKDD 2009. 2009. p. 426-438 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 5476 LNAI).

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

TY - GEN

T1 - RV-SVM

T2 - An efficient method for learning ranking SVM

AU - Hwanjo, Yu

AU - Youngdae, Kim

AU - Seungwon, Hwang

PY - 2009/7/23

Y1 - 2009/7/23

N2 - Learning ranking (or preference) functions has become an important data mining task in recent years, as various applications have been found in information retrieval. Among rank learning methods, ranking SVMhas been favorably applied to various applications, e.g., optimizing search engines, improving data retrieval quality. In this paper, we first develop a 1-norm ranking SVM that is faster in testing than the standard ranking SVM, and propose Ranking Vector SVM (RV-SVM) that revises the 1-norm ranking SVM for faster training. The number of variables in the RV-SVM is significantly smaller, thus the RV-SVM trains much faster than the other ranking SVMs.We experimentally compared the RV-SVM with the state-of-the-art rank learning method provided in SVM-light. The RV-SVMuses much less support vectors and trains much faster for nonlinear kernels than the SVM-light. The accuracies of RV-SVM and SVM-light are comparable on relatively large data sets. Our implementation of RV-SVM is posted at http://iis.postech.ac.kr/rv-svm.

AB - Learning ranking (or preference) functions has become an important data mining task in recent years, as various applications have been found in information retrieval. Among rank learning methods, ranking SVMhas been favorably applied to various applications, e.g., optimizing search engines, improving data retrieval quality. In this paper, we first develop a 1-norm ranking SVM that is faster in testing than the standard ranking SVM, and propose Ranking Vector SVM (RV-SVM) that revises the 1-norm ranking SVM for faster training. The number of variables in the RV-SVM is significantly smaller, thus the RV-SVM trains much faster than the other ranking SVMs.We experimentally compared the RV-SVM with the state-of-the-art rank learning method provided in SVM-light. The RV-SVMuses much less support vectors and trains much faster for nonlinear kernels than the SVM-light. The accuracies of RV-SVM and SVM-light are comparable on relatively large data sets. Our implementation of RV-SVM is posted at http://iis.postech.ac.kr/rv-svm.

UR - http://www.scopus.com/inward/record.url?scp=67650686496&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=67650686496&partnerID=8YFLogxK

U2 - 10.1007/978-3-642-01307-2_39

DO - 10.1007/978-3-642-01307-2_39

M3 - Conference contribution

AN - SCOPUS:67650686496

SN - 3642013066

SN - 9783642013065

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

SP - 426

EP - 438

BT - 13th Pacific-Asia Conference on Knowledge Discovery and Data Mining, PAKDD 2009

ER -

Hwanjo Y, Youngdae K, Seungwon H. RV-SVM: An efficient method for learning ranking SVM. In 13th Pacific-Asia Conference on Knowledge Discovery and Data Mining, PAKDD 2009. 2009. p. 426-438. (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)). https://doi.org/10.1007/978-3-642-01307-2_39