Modeling Documents as Mixtures of Persons for Expert Finding

Pavel Serdyukov; Djoerd Hiemstra

doi:10.1007/978-3-540-78646-7_29

Modeling Documents as Mixtures of Persons for Expert Finding

Pavel Serdyukov, Djoerd Hiemstra

Databases (Former)

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

53 Citations (Scopus)

219 Downloads (Pure)

Abstract

In this paper we address the problem of searching for knowledgeable persons within the enterprise, known as the expert finding (or expert search) task. We present a probabilistic algorithm using the assumption that terms in documents are produced by people who are mentioned in them.We represent documents retrieved to a query as mixtures of candidate experts language models. Two methods of personal language models extraction are proposed, as well as the way of combining them with other evidences of expertise. Experiments conducted with the TREC Enterprise collection demonstrate the superiority of our approach in comparison with the best one among existing solutions.

Original language	Undefined
Title of host publication	Proceedings of the 30th European Conference on Information Retrieval (ECIR 2008)
Place of Publication	London
Publisher	Springer
Pages	309-320
Number of pages	12
ISBN (Print)	978-3-540-78645-0
DOIs	https://doi.org/10.1007/978-3-540-78646-7_29
Publication status	Published - 27 Mar 2008
Event	30th European Conference on Information Retrieval, ECIR 2008: (IR Resarch) - Glasgow, United Kingdom Duration: 31 Mar 2008 → 4 Apr 2008 Conference number: 30

Publication series

Name	Lecture Notes in Computer Science
Publisher	Springer Verlag
Number	1
Volume	4956

Conference

Conference	30th European Conference on Information Retrieval, ECIR 2008
Abbreviated title	ECIR
Country/Territory	United Kingdom
City	Glasgow
Period	31/03/08 → 4/04/08

Keywords

DB-IR: INFORMATION RETRIEVAL
IR-64731
METIS-250964
EWI-12311

Access to Document

10.1007/978-3-540-78646-7_29

Cite this

@inproceedings{b5430ec9826944c3b624373c2899fa47,

title = "Modeling Documents as Mixtures of Persons for Expert Finding",

abstract = "In this paper we address the problem of searching for knowledgeable persons within the enterprise, known as the expert finding (or expert search) task. We present a probabilistic algorithm using the assumption that terms in documents are produced by people who are mentioned in them.We represent documents retrieved to a query as mixtures of candidate experts language models. Two methods of personal language models extraction are proposed, as well as the way of combining them with other evidences of expertise. Experiments conducted with the TREC Enterprise collection demonstrate the superiority of our approach in comparison with the best one among existing solutions.",

keywords = "DB-IR: INFORMATION RETRIEVAL, IR-64731, METIS-250964, EWI-12311",

author = "Pavel Serdyukov and Djoerd Hiemstra",

note = "Best student paper award; 30th European Conference on Information Retrieval, ECIR 2008 ; Conference date: 31-03-2008 Through 04-04-2008",

year = "2008",

month = mar,

day = "27",

doi = "10.1007/978-3-540-78646-7_29",

language = "Undefined",

isbn = "978-3-540-78645-0",

series = "Lecture Notes in Computer Science",

publisher = "Springer",

number = "1",

pages = "309--320",

booktitle = "Proceedings of the 30th European Conference on Information Retrieval (ECIR 2008)",

address = "Germany",

}

Serdyukov, P & Hiemstra, D 2008, Modeling Documents as Mixtures of Persons for Expert Finding. in Proceedings of the 30th European Conference on Information Retrieval (ECIR 2008)., 10.1007/978-3-540-78646-7_29, Lecture Notes in Computer Science, no. 1, vol. 4956, Springer, London, pp. 309-320, 30th European Conference on Information Retrieval, ECIR 2008, Glasgow, United Kingdom, 31/03/08. https://doi.org/10.1007/978-3-540-78646-7_29

Modeling Documents as Mixtures of Persons for Expert Finding. / Serdyukov, Pavel; Hiemstra, Djoerd.
Proceedings of the 30th European Conference on Information Retrieval (ECIR 2008). London: Springer, 2008. p. 309-320 10.1007/978-3-540-78646-7_29 (Lecture Notes in Computer Science; Vol. 4956, No. 1).

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

TY - GEN

T1 - Modeling Documents as Mixtures of Persons for Expert Finding

AU - Serdyukov, Pavel

AU - Hiemstra, Djoerd

N1 - Conference code: 30

PY - 2008/3/27

Y1 - 2008/3/27

N2 - In this paper we address the problem of searching for knowledgeable persons within the enterprise, known as the expert finding (or expert search) task. We present a probabilistic algorithm using the assumption that terms in documents are produced by people who are mentioned in them.We represent documents retrieved to a query as mixtures of candidate experts language models. Two methods of personal language models extraction are proposed, as well as the way of combining them with other evidences of expertise. Experiments conducted with the TREC Enterprise collection demonstrate the superiority of our approach in comparison with the best one among existing solutions.

AB - In this paper we address the problem of searching for knowledgeable persons within the enterprise, known as the expert finding (or expert search) task. We present a probabilistic algorithm using the assumption that terms in documents are produced by people who are mentioned in them.We represent documents retrieved to a query as mixtures of candidate experts language models. Two methods of personal language models extraction are proposed, as well as the way of combining them with other evidences of expertise. Experiments conducted with the TREC Enterprise collection demonstrate the superiority of our approach in comparison with the best one among existing solutions.

KW - DB-IR: INFORMATION RETRIEVAL

KW - IR-64731

KW - METIS-250964

KW - EWI-12311

U2 - 10.1007/978-3-540-78646-7_29

DO - 10.1007/978-3-540-78646-7_29

M3 - Conference contribution

SN - 978-3-540-78645-0

T3 - Lecture Notes in Computer Science

SP - 309

EP - 320

BT - Proceedings of the 30th European Conference on Information Retrieval (ECIR 2008)

PB - Springer

CY - London

T2 - 30th European Conference on Information Retrieval, ECIR 2008

Y2 - 31 March 2008 through 4 April 2008

ER -