Improved Relevance feedback using density-based clustering

Kareem Darwish, Ahmed El Deeb, Noha Yousri, Mohamed S. Kamel

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

Abstract

Relevance feedback (RFB) involves requesting some user judgments for an initial set of search results and then using these judgments to improve search results. Typical queries may have multiple possible interpretations or facets, only one of which is relevant to a user's need, but top search results may be dominated by one interpretation or facet. Thus, if the user is only given the top results to inspect, none of them may be relevant. One way to solve this is to intentionally diversify the top few results to cover multiple interpretations. This paper proposes the use of density-based clustering for the purpose of results diversification in the context of RFB. Other traditional clustering algorithms are also used for a comparative study. Clustering is compared to a baseline that nominates the top results from an initial ranked list and compared to using the top results after re-ranking using Maximal Marginal Relevance. The results show that density-based clustering achieves the best results with a statistically significant improvement of 12% over the baseline.

Original languageEnglish
Title of host publicationProceedings of the 2010 10th International Conference on Intelligent Systems Design and Applications, ISDA'10
Pages580-585
Number of pages6
DOIs
Publication statusPublished - 1 Dec 2010
Externally publishedYes
Event2010 10th International Conference on Intelligent Systems Design and Applications, ISDA'10 - Cairo, Egypt
Duration: 29 Nov 20101 Dec 2010

Other

Other2010 10th International Conference on Intelligent Systems Design and Applications, ISDA'10
CountryEgypt
CityCairo
Period29/11/101/12/10

Fingerprint

Feedback
Clustering algorithms

Keywords

  • Clustering
  • Query reformulation
  • Relevance feedback
  • Selection process

ASJC Scopus subject areas

  • Artificial Intelligence
  • Computer Science Applications
  • Hardware and Architecture

Cite this

Darwish, K., El Deeb, A., Yousri, N., & Kamel, M. S. (2010). Improved Relevance feedback using density-based clustering. In Proceedings of the 2010 10th International Conference on Intelligent Systems Design and Applications, ISDA'10 (pp. 580-585). [5687203] https://doi.org/10.1109/ISDA.2010.5687203

Improved Relevance feedback using density-based clustering. / Darwish, Kareem; El Deeb, Ahmed; Yousri, Noha; Kamel, Mohamed S.

Proceedings of the 2010 10th International Conference on Intelligent Systems Design and Applications, ISDA'10. 2010. p. 580-585 5687203.

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

Darwish, K, El Deeb, A, Yousri, N & Kamel, MS 2010, Improved Relevance feedback using density-based clustering. in Proceedings of the 2010 10th International Conference on Intelligent Systems Design and Applications, ISDA'10., 5687203, pp. 580-585, 2010 10th International Conference on Intelligent Systems Design and Applications, ISDA'10, Cairo, Egypt, 29/11/10. https://doi.org/10.1109/ISDA.2010.5687203
Darwish K, El Deeb A, Yousri N, Kamel MS. Improved Relevance feedback using density-based clustering. In Proceedings of the 2010 10th International Conference on Intelligent Systems Design and Applications, ISDA'10. 2010. p. 580-585. 5687203 https://doi.org/10.1109/ISDA.2010.5687203
Darwish, Kareem ; El Deeb, Ahmed ; Yousri, Noha ; Kamel, Mohamed S. / Improved Relevance feedback using density-based clustering. Proceedings of the 2010 10th International Conference on Intelligent Systems Design and Applications, ISDA'10. 2010. pp. 580-585
@inproceedings{30259180a1644475ba19c01581ff3a7f,
title = "Improved Relevance feedback using density-based clustering",
abstract = "Relevance feedback (RFB) involves requesting some user judgments for an initial set of search results and then using these judgments to improve search results. Typical queries may have multiple possible interpretations or facets, only one of which is relevant to a user's need, but top search results may be dominated by one interpretation or facet. Thus, if the user is only given the top results to inspect, none of them may be relevant. One way to solve this is to intentionally diversify the top few results to cover multiple interpretations. This paper proposes the use of density-based clustering for the purpose of results diversification in the context of RFB. Other traditional clustering algorithms are also used for a comparative study. Clustering is compared to a baseline that nominates the top results from an initial ranked list and compared to using the top results after re-ranking using Maximal Marginal Relevance. The results show that density-based clustering achieves the best results with a statistically significant improvement of 12{\%} over the baseline.",
keywords = "Clustering, Query reformulation, Relevance feedback, Selection process",
author = "Kareem Darwish and {El Deeb}, Ahmed and Noha Yousri and Kamel, {Mohamed S.}",
year = "2010",
month = "12",
day = "1",
doi = "10.1109/ISDA.2010.5687203",
language = "English",
isbn = "9781424481354",
pages = "580--585",
booktitle = "Proceedings of the 2010 10th International Conference on Intelligent Systems Design and Applications, ISDA'10",

}

TY - GEN

T1 - Improved Relevance feedback using density-based clustering

AU - Darwish, Kareem

AU - El Deeb, Ahmed

AU - Yousri, Noha

AU - Kamel, Mohamed S.

PY - 2010/12/1

Y1 - 2010/12/1

N2 - Relevance feedback (RFB) involves requesting some user judgments for an initial set of search results and then using these judgments to improve search results. Typical queries may have multiple possible interpretations or facets, only one of which is relevant to a user's need, but top search results may be dominated by one interpretation or facet. Thus, if the user is only given the top results to inspect, none of them may be relevant. One way to solve this is to intentionally diversify the top few results to cover multiple interpretations. This paper proposes the use of density-based clustering for the purpose of results diversification in the context of RFB. Other traditional clustering algorithms are also used for a comparative study. Clustering is compared to a baseline that nominates the top results from an initial ranked list and compared to using the top results after re-ranking using Maximal Marginal Relevance. The results show that density-based clustering achieves the best results with a statistically significant improvement of 12% over the baseline.

AB - Relevance feedback (RFB) involves requesting some user judgments for an initial set of search results and then using these judgments to improve search results. Typical queries may have multiple possible interpretations or facets, only one of which is relevant to a user's need, but top search results may be dominated by one interpretation or facet. Thus, if the user is only given the top results to inspect, none of them may be relevant. One way to solve this is to intentionally diversify the top few results to cover multiple interpretations. This paper proposes the use of density-based clustering for the purpose of results diversification in the context of RFB. Other traditional clustering algorithms are also used for a comparative study. Clustering is compared to a baseline that nominates the top results from an initial ranked list and compared to using the top results after re-ranking using Maximal Marginal Relevance. The results show that density-based clustering achieves the best results with a statistically significant improvement of 12% over the baseline.

KW - Clustering

KW - Query reformulation

KW - Relevance feedback

KW - Selection process

UR - http://www.scopus.com/inward/record.url?scp=79851480795&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=79851480795&partnerID=8YFLogxK

U2 - 10.1109/ISDA.2010.5687203

DO - 10.1109/ISDA.2010.5687203

M3 - Conference contribution

AN - SCOPUS:79851480795

SN - 9781424481354

SP - 580

EP - 585

BT - Proceedings of the 2010 10th International Conference on Intelligent Systems Design and Applications, ISDA'10

ER -