Measuring the likelihood property of scoring functions in general retrieval models

R. Bache, M. Baillie, F. Crestani

Research output: Contribution to journalArticle

1 Citation (Scopus)

Abstract

Although retrieval systems based on probabilistic models will rank the objects (e.g., documents) being retrieved according to the probability of some matching criterion (e.g., relevance), they rarely yield an actual probability, and the scoring function is interpreted to be purely ordinal within a given retrieval task. In this brief communication, it is shown that some scoring functions possess the likelihood property, which means that the scoring function indicates the likelihood of matching when compared to other retrieval tasks, which is potentially more useful than pure ranking although it cannot be interpreted as an actual probability. This property can be detected by using two modified effectiveness measures: entire precision and entire recall.
LanguageEnglish
Pages1294-1297
Number of pages4
JournalJournal of the American Society for Information Science and Technology
Volume60
Issue number6
DOIs
Publication statusPublished - 2009

Fingerprint

Communication
Statistical Models

Keywords

  • information retrieval
  • probability
  • probabilistic models
  • retrieval systems

Cite this

@article{5a0d081f372b41bdb4938eb059fbdf04,
title = "Measuring the likelihood property of scoring functions in general retrieval models",
abstract = "Although retrieval systems based on probabilistic models will rank the objects (e.g., documents) being retrieved according to the probability of some matching criterion (e.g., relevance), they rarely yield an actual probability, and the scoring function is interpreted to be purely ordinal within a given retrieval task. In this brief communication, it is shown that some scoring functions possess the likelihood property, which means that the scoring function indicates the likelihood of matching when compared to other retrieval tasks, which is potentially more useful than pure ranking although it cannot be interpreted as an actual probability. This property can be detected by using two modified effectiveness measures: entire precision and entire recall.",
keywords = "information retrieval, probability, probabilistic models, retrieval systems",
author = "R. Bache and M. Baillie and F. Crestani",
year = "2009",
doi = "10.1002/asi.21048",
language = "English",
volume = "60",
pages = "1294--1297",
journal = "Journal of the American Society for Information Science and Technology",
issn = "1532-2890",
number = "6",

}

Measuring the likelihood property of scoring functions in general retrieval models. / Bache, R.; Baillie, M.; Crestani, F.

In: Journal of the American Society for Information Science and Technology, Vol. 60, No. 6, 2009, p. 1294-1297.

Research output: Contribution to journalArticle

TY - JOUR

T1 - Measuring the likelihood property of scoring functions in general retrieval models

AU - Bache, R.

AU - Baillie, M.

AU - Crestani, F.

PY - 2009

Y1 - 2009

N2 - Although retrieval systems based on probabilistic models will rank the objects (e.g., documents) being retrieved according to the probability of some matching criterion (e.g., relevance), they rarely yield an actual probability, and the scoring function is interpreted to be purely ordinal within a given retrieval task. In this brief communication, it is shown that some scoring functions possess the likelihood property, which means that the scoring function indicates the likelihood of matching when compared to other retrieval tasks, which is potentially more useful than pure ranking although it cannot be interpreted as an actual probability. This property can be detected by using two modified effectiveness measures: entire precision and entire recall.

AB - Although retrieval systems based on probabilistic models will rank the objects (e.g., documents) being retrieved according to the probability of some matching criterion (e.g., relevance), they rarely yield an actual probability, and the scoring function is interpreted to be purely ordinal within a given retrieval task. In this brief communication, it is shown that some scoring functions possess the likelihood property, which means that the scoring function indicates the likelihood of matching when compared to other retrieval tasks, which is potentially more useful than pure ranking although it cannot be interpreted as an actual probability. This property can be detected by using two modified effectiveness measures: entire precision and entire recall.

KW - information retrieval

KW - probability

KW - probabilistic models

KW - retrieval systems

U2 - 10.1002/asi.21048

DO - 10.1002/asi.21048

M3 - Article

VL - 60

SP - 1294

EP - 1297

JO - Journal of the American Society for Information Science and Technology

T2 - Journal of the American Society for Information Science and Technology

JF - Journal of the American Society for Information Science and Technology

SN - 1532-2890

IS - 6

ER -