Modeling Multi-step Relevance Propagation for Expert Finding

Pavel Serdyukov, H. Rode, Djoerd Hiemstra

Research output: Chapter in Book/Report/Conference proceedingConference contributionAcademicpeer-review

70 Citations (Scopus)
60 Downloads (Pure)

Abstract

An expert finding system allows a user to type a simple text query and retrieve names and contact information of individuals that possess the expertise expressed in the query. This paper proposes a novel approach to expert finding in large enterprises or intranets by modeling candidate experts (persons), web documents and various relations among them with so-called expertise graphs. As distinct from the state-of-the-art approaches estimating personal expertise through one-step propagation of relevance probability from documents to the related candidates, our methods are based on the principle of multi-step relevance propagation in topic-specific expertise graphs. We model the process of expert finding by probabilistic random walks of three kinds: finite, infinite and absorbing. Experiments on TREC Enterprise Track data originating from two large organizations show that our methods using multi-step relevance propagation improve over the baseline one-step propagation based method in almost all cases.
Original languageUndefined
Title of host publicationProceeding of the 17th ACM Conference on Information and Knowledge Management (CIKM2008)
Place of PublicationNew York
PublisherAssociation for Computing Machinery (ACM)
Pages1133-1142
Number of pages10
ISBN (Print)978-1-59593-991-3
DOIs
Publication statusPublished - Oct 2008
Event17th ACM Conference on Information and Knowledge Management, CIKM 2008 - Napa Valley, United States
Duration: 26 Oct 200830 Oct 2008
Conference number: 17

Publication series

Name
NumberSupplement

Conference

Conference17th ACM Conference on Information and Knowledge Management, CIKM 2008
Abbreviated titleCIKM
CountryUnited States
CityNapa Valley
Period26/10/0830/10/08

Keywords

  • EWI-13455
  • CR-H.3
  • METIS-251189
  • DB-IR: INFORMATION RETRIEVAL
  • CR-H.3.3

Cite this

Serdyukov, P., Rode, H., & Hiemstra, D. (2008). Modeling Multi-step Relevance Propagation for Expert Finding. In Proceeding of the 17th ACM Conference on Information and Knowledge Management (CIKM2008) (pp. 1133-1142). New York: Association for Computing Machinery (ACM). https://doi.org/10.1145/1458082.1458232
Serdyukov, Pavel ; Rode, H. ; Hiemstra, Djoerd. / Modeling Multi-step Relevance Propagation for Expert Finding. Proceeding of the 17th ACM Conference on Information and Knowledge Management (CIKM2008). New York : Association for Computing Machinery (ACM), 2008. pp. 1133-1142
@inproceedings{490e5273e13143368606055735592c6c,
title = "Modeling Multi-step Relevance Propagation for Expert Finding",
abstract = "An expert finding system allows a user to type a simple text query and retrieve names and contact information of individuals that possess the expertise expressed in the query. This paper proposes a novel approach to expert finding in large enterprises or intranets by modeling candidate experts (persons), web documents and various relations among them with so-called expertise graphs. As distinct from the state-of-the-art approaches estimating personal expertise through one-step propagation of relevance probability from documents to the related candidates, our methods are based on the principle of multi-step relevance propagation in topic-specific expertise graphs. We model the process of expert finding by probabilistic random walks of three kinds: finite, infinite and absorbing. Experiments on TREC Enterprise Track data originating from two large organizations show that our methods using multi-step relevance propagation improve over the baseline one-step propagation based method in almost all cases.",
keywords = "EWI-13455, CR-H.3, METIS-251189, DB-IR: INFORMATION RETRIEVAL, CR-H.3.3",
author = "Pavel Serdyukov and H. Rode and Djoerd Hiemstra",
year = "2008",
month = "10",
doi = "10.1145/1458082.1458232",
language = "Undefined",
isbn = "978-1-59593-991-3",
publisher = "Association for Computing Machinery (ACM)",
number = "Supplement",
pages = "1133--1142",
booktitle = "Proceeding of the 17th ACM Conference on Information and Knowledge Management (CIKM2008)",
address = "United States",

}

Serdyukov, P, Rode, H & Hiemstra, D 2008, Modeling Multi-step Relevance Propagation for Expert Finding. in Proceeding of the 17th ACM Conference on Information and Knowledge Management (CIKM2008). Association for Computing Machinery (ACM), New York, pp. 1133-1142, 17th ACM Conference on Information and Knowledge Management, CIKM 2008, Napa Valley, United States, 26/10/08. https://doi.org/10.1145/1458082.1458232

Modeling Multi-step Relevance Propagation for Expert Finding. / Serdyukov, Pavel; Rode, H.; Hiemstra, Djoerd.

Proceeding of the 17th ACM Conference on Information and Knowledge Management (CIKM2008). New York : Association for Computing Machinery (ACM), 2008. p. 1133-1142.

Research output: Chapter in Book/Report/Conference proceedingConference contributionAcademicpeer-review

TY - GEN

T1 - Modeling Multi-step Relevance Propagation for Expert Finding

AU - Serdyukov, Pavel

AU - Rode, H.

AU - Hiemstra, Djoerd

PY - 2008/10

Y1 - 2008/10

N2 - An expert finding system allows a user to type a simple text query and retrieve names and contact information of individuals that possess the expertise expressed in the query. This paper proposes a novel approach to expert finding in large enterprises or intranets by modeling candidate experts (persons), web documents and various relations among them with so-called expertise graphs. As distinct from the state-of-the-art approaches estimating personal expertise through one-step propagation of relevance probability from documents to the related candidates, our methods are based on the principle of multi-step relevance propagation in topic-specific expertise graphs. We model the process of expert finding by probabilistic random walks of three kinds: finite, infinite and absorbing. Experiments on TREC Enterprise Track data originating from two large organizations show that our methods using multi-step relevance propagation improve over the baseline one-step propagation based method in almost all cases.

AB - An expert finding system allows a user to type a simple text query and retrieve names and contact information of individuals that possess the expertise expressed in the query. This paper proposes a novel approach to expert finding in large enterprises or intranets by modeling candidate experts (persons), web documents and various relations among them with so-called expertise graphs. As distinct from the state-of-the-art approaches estimating personal expertise through one-step propagation of relevance probability from documents to the related candidates, our methods are based on the principle of multi-step relevance propagation in topic-specific expertise graphs. We model the process of expert finding by probabilistic random walks of three kinds: finite, infinite and absorbing. Experiments on TREC Enterprise Track data originating from two large organizations show that our methods using multi-step relevance propagation improve over the baseline one-step propagation based method in almost all cases.

KW - EWI-13455

KW - CR-H.3

KW - METIS-251189

KW - DB-IR: INFORMATION RETRIEVAL

KW - CR-H.3.3

U2 - 10.1145/1458082.1458232

DO - 10.1145/1458082.1458232

M3 - Conference contribution

SN - 978-1-59593-991-3

SP - 1133

EP - 1142

BT - Proceeding of the 17th ACM Conference on Information and Knowledge Management (CIKM2008)

PB - Association for Computing Machinery (ACM)

CY - New York

ER -

Serdyukov P, Rode H, Hiemstra D. Modeling Multi-step Relevance Propagation for Expert Finding. In Proceeding of the 17th ACM Conference on Information and Knowledge Management (CIKM2008). New York: Association for Computing Machinery (ACM). 2008. p. 1133-1142 https://doi.org/10.1145/1458082.1458232