Insights from viewing ranked retrieval as rank aggregation

Holger Bast, Ingmar Weber

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

4 Citations (Scopus)

Abstract

We view a variety of established methods for ranked retrieval from a common angle, namely as a process of combining query-independent rankings that were precomputed for certain attributes. Apart from a general insight into what effectively distinguishes various schemes from each other, we obtain three specific results concerned with conceptbased retrieval. First, we prove that latent semantic indexing (LSI) can be implemented to answer queries in time proportional to the number of words in the query, which improves over the standard implementation by an order of magnitude; a similar result is established for LSI's probabilistic sibling PLSI. Second, we give a simple and precise characterization of the extent, to which latent semantic indexing (LSI) can deal with polysems, and when it fails to do so. Third, we demonstrate that the recombination of the intricate, yet relatively cheap mechanism of PLSI for mapping queries to attributes, with a simplistic, easy-to-compute set of document rankings gives a retrieval performance which is at least as good as that of the most sophisticated concept-based retrieval schemes and which does not require any pre-computation.

Original languageEnglish
Title of host publicationProceedings - International Workshop on Challenges in Web Information Retrieval and Integration, WIRI'05
Pages232-237
Number of pages6
Volume2005
DOIs
Publication statusPublished - 1 Dec 2005
Externally publishedYes
EventInternational Workshop on Challenges in Web Information Retrieval and Integration, WIRI'05 - Tokyo, Japan
Duration: 8 Apr 20059 Apr 2005

Other

OtherInternational Workshop on Challenges in Web Information Retrieval and Integration, WIRI'05
CountryJapan
CityTokyo
Period8/4/059/4/05

Fingerprint

Agglomeration
Semantics

ASJC Scopus subject areas

  • Engineering(all)

Cite this

Bast, H., & Weber, I. (2005). Insights from viewing ranked retrieval as rank aggregation. In Proceedings - International Workshop on Challenges in Web Information Retrieval and Integration, WIRI'05 (Vol. 2005, pp. 232-237). [1553019] https://doi.org/10.1109/WIRI.2005.19

Insights from viewing ranked retrieval as rank aggregation. / Bast, Holger; Weber, Ingmar.

Proceedings - International Workshop on Challenges in Web Information Retrieval and Integration, WIRI'05. Vol. 2005 2005. p. 232-237 1553019.

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

Bast, H & Weber, I 2005, Insights from viewing ranked retrieval as rank aggregation. in Proceedings - International Workshop on Challenges in Web Information Retrieval and Integration, WIRI'05. vol. 2005, 1553019, pp. 232-237, International Workshop on Challenges in Web Information Retrieval and Integration, WIRI'05, Tokyo, Japan, 8/4/05. https://doi.org/10.1109/WIRI.2005.19
Bast H, Weber I. Insights from viewing ranked retrieval as rank aggregation. In Proceedings - International Workshop on Challenges in Web Information Retrieval and Integration, WIRI'05. Vol. 2005. 2005. p. 232-237. 1553019 https://doi.org/10.1109/WIRI.2005.19
Bast, Holger ; Weber, Ingmar. / Insights from viewing ranked retrieval as rank aggregation. Proceedings - International Workshop on Challenges in Web Information Retrieval and Integration, WIRI'05. Vol. 2005 2005. pp. 232-237
@inproceedings{3147dc62009445e29a5414d772253e3b,
title = "Insights from viewing ranked retrieval as rank aggregation",
abstract = "We view a variety of established methods for ranked retrieval from a common angle, namely as a process of combining query-independent rankings that were precomputed for certain attributes. Apart from a general insight into what effectively distinguishes various schemes from each other, we obtain three specific results concerned with conceptbased retrieval. First, we prove that latent semantic indexing (LSI) can be implemented to answer queries in time proportional to the number of words in the query, which improves over the standard implementation by an order of magnitude; a similar result is established for LSI's probabilistic sibling PLSI. Second, we give a simple and precise characterization of the extent, to which latent semantic indexing (LSI) can deal with polysems, and when it fails to do so. Third, we demonstrate that the recombination of the intricate, yet relatively cheap mechanism of PLSI for mapping queries to attributes, with a simplistic, easy-to-compute set of document rankings gives a retrieval performance which is at least as good as that of the most sophisticated concept-based retrieval schemes and which does not require any pre-computation.",
author = "Holger Bast and Ingmar Weber",
year = "2005",
month = "12",
day = "1",
doi = "10.1109/WIRI.2005.19",
language = "English",
isbn = "0769524141",
volume = "2005",
pages = "232--237",
booktitle = "Proceedings - International Workshop on Challenges in Web Information Retrieval and Integration, WIRI'05",

}

TY - GEN

T1 - Insights from viewing ranked retrieval as rank aggregation

AU - Bast, Holger

AU - Weber, Ingmar

PY - 2005/12/1

Y1 - 2005/12/1

N2 - We view a variety of established methods for ranked retrieval from a common angle, namely as a process of combining query-independent rankings that were precomputed for certain attributes. Apart from a general insight into what effectively distinguishes various schemes from each other, we obtain three specific results concerned with conceptbased retrieval. First, we prove that latent semantic indexing (LSI) can be implemented to answer queries in time proportional to the number of words in the query, which improves over the standard implementation by an order of magnitude; a similar result is established for LSI's probabilistic sibling PLSI. Second, we give a simple and precise characterization of the extent, to which latent semantic indexing (LSI) can deal with polysems, and when it fails to do so. Third, we demonstrate that the recombination of the intricate, yet relatively cheap mechanism of PLSI for mapping queries to attributes, with a simplistic, easy-to-compute set of document rankings gives a retrieval performance which is at least as good as that of the most sophisticated concept-based retrieval schemes and which does not require any pre-computation.

AB - We view a variety of established methods for ranked retrieval from a common angle, namely as a process of combining query-independent rankings that were precomputed for certain attributes. Apart from a general insight into what effectively distinguishes various schemes from each other, we obtain three specific results concerned with conceptbased retrieval. First, we prove that latent semantic indexing (LSI) can be implemented to answer queries in time proportional to the number of words in the query, which improves over the standard implementation by an order of magnitude; a similar result is established for LSI's probabilistic sibling PLSI. Second, we give a simple and precise characterization of the extent, to which latent semantic indexing (LSI) can deal with polysems, and when it fails to do so. Third, we demonstrate that the recombination of the intricate, yet relatively cheap mechanism of PLSI for mapping queries to attributes, with a simplistic, easy-to-compute set of document rankings gives a retrieval performance which is at least as good as that of the most sophisticated concept-based retrieval schemes and which does not require any pre-computation.

UR - http://www.scopus.com/inward/record.url?scp=33845364011&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=33845364011&partnerID=8YFLogxK

U2 - 10.1109/WIRI.2005.19

DO - 10.1109/WIRI.2005.19

M3 - Conference contribution

SN - 0769524141

SN - 9780769524146

VL - 2005

SP - 232

EP - 237

BT - Proceedings - International Workshop on Challenges in Web Information Retrieval and Integration, WIRI'05

ER -