Ensemble clustering for result diversification

Dong-Phuong Nguyen; Djoerd Hiemstra

Ensemble clustering for result diversification

Dong-Phuong Nguyen, Djoerd Hiemstra

Databases (Former)

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › Academic

74 Downloads (Pure)

Abstract

This paper describes the participation of the University of Twente in the Web track of TREC 2012. Our baseline approach uses the Mirex toolkit, an open source tool that sequantially scans all the documents. For result diversification, we experimented with improving the quality of clusters through ensemble clustering. We combined clusters obtained by different clustering methods (such as LDA and K-means) and clusters obtained by using different types of data (such as document text and anchor text). Our two-layer ensemble run performed better than the LDA based diversification and also better than a non-diversification run.

Original language	Undefined
Title of host publication	Proceedings of the Twenty First Text REtrieval Conference (TREC 2012)
Place of Publication	Gaithersburg, MD, USA
Publisher	NIST
Pages	83
Number of pages	4
ISBN (Print)	not assigned
Publication status	Published - 6 Nov 2012
Event	Twenty-First Text REtrieval Conference, TREC-21 2012 - Gaithersburg, United States Duration: 6 Nov 2012 → 9 Nov 2012 Conference number: 21

Publication series

Name	NIST Special Publications
Publisher	NIST
Number	SP 500-298
Volume	SP 500-298

Workshop

Workshop	Twenty-First Text REtrieval Conference, TREC-21 2012
Abbreviated title	TREC
Country/Territory	United States
City	Gaithersburg
Period	6/11/12 → 9/11/12

Keywords

EWI-22911
IR-84255
METIS-296227

Access to Document

http://trec.nist.gov/pubs/trec21/papers/utwente.web.final.pdf

Cite this

@inproceedings{4de4213760954558819f1912cacc3a38,

title = "Ensemble clustering for result diversification",

abstract = "This paper describes the participation of the University of Twente in the Web track of TREC 2012. Our baseline approach uses the Mirex toolkit, an open source tool that sequantially scans all the documents. For result diversification, we experimented with improving the quality of clusters through ensemble clustering. We combined clusters obtained by different clustering methods (such as LDA and K-means) and clusters obtained by using different types of data (such as document text and anchor text). Our two-layer ensemble run performed better than the LDA based diversification and also better than a non-diversification run.",

keywords = "EWI-22911, IR-84255, METIS-296227",

author = "Dong-Phuong Nguyen and Djoerd Hiemstra",

year = "2012",

month = nov,

day = "6",

language = "Undefined",

isbn = "not assigned",

series = "NIST Special Publications",

publisher = "NIST",

number = "SP 500-298",

pages = "83",

booktitle = "Proceedings of the Twenty First Text REtrieval Conference (TREC 2012)",

note = "Twenty-First Text REtrieval Conference, TREC-21 2012 ; Conference date: 06-11-2012 Through 09-11-2012",

}

Nguyen, D-P & Hiemstra, D 2012, Ensemble clustering for result diversification. in Proceedings of the Twenty First Text REtrieval Conference (TREC 2012). NIST Special Publications, no. SP 500-298, vol. SP 500-298, NIST, Gaithersburg, MD, USA, pp. 83, Twenty-First Text REtrieval Conference, TREC-21 2012, Gaithersburg, Maryland, United States, 6/11/12. <http://trec.nist.gov/pubs/trec21/papers/utwente.web.final.pdf>

TY - GEN

T1 - Ensemble clustering for result diversification

AU - Nguyen, Dong-Phuong

AU - Hiemstra, Djoerd

N1 - Conference code: 21

PY - 2012/11/6

Y1 - 2012/11/6

N2 - This paper describes the participation of the University of Twente in the Web track of TREC 2012. Our baseline approach uses the Mirex toolkit, an open source tool that sequantially scans all the documents. For result diversification, we experimented with improving the quality of clusters through ensemble clustering. We combined clusters obtained by different clustering methods (such as LDA and K-means) and clusters obtained by using different types of data (such as document text and anchor text). Our two-layer ensemble run performed better than the LDA based diversification and also better than a non-diversification run.

AB - This paper describes the participation of the University of Twente in the Web track of TREC 2012. Our baseline approach uses the Mirex toolkit, an open source tool that sequantially scans all the documents. For result diversification, we experimented with improving the quality of clusters through ensemble clustering. We combined clusters obtained by different clustering methods (such as LDA and K-means) and clusters obtained by using different types of data (such as document text and anchor text). Our two-layer ensemble run performed better than the LDA based diversification and also better than a non-diversification run.

KW - EWI-22911

KW - IR-84255

KW - METIS-296227

M3 - Conference contribution

SN - not assigned

T3 - NIST Special Publications

SP - 83

BT - Proceedings of the Twenty First Text REtrieval Conference (TREC 2012)

PB - NIST

CY - Gaithersburg, MD, USA

T2 - Twenty-First Text REtrieval Conference, TREC-21 2012

Y2 - 6 November 2012 through 9 November 2012

ER -