On generalizing the Two‐Poisson Model

Padmini Srinivasan

Journal of the American Society for Information Science, 1990, vol. 41, issue 1, 61-66

Abstract: It is well recognized that automatic indexing is one of the important functions of a modern Document Retrieval System. Numerous techniques for this function have been proposed in the literature ranging from purely statistical to linguistically complex mechanisms. Most of these techniques result from examining properties of terms. In this article term distribution is examined within the framework of the Poisson Models. Specifically the effectiveness of the Two‐Poisson and the Three‐Poisson Models are examined. The more general Three‐Poisson model is examined to see if generalization results in increased effectiveness. The results show that the Two‐Poisson model is only moderately effective in identifying index terms. In addition, generalization to the Three‐Poisson does not give any additional power. The only model within the framework of Poisson models which consistently works well is the basic One‐Poisson model. The article concludes with a discussion on term distribution information. © 1990 John Wiley & Sons, Inc.

Date: 1990
References: Add references at CitEc
Citations:

Downloads: (external link)
https://doi.org/10.1002/(SICI)1097-4571(199001)41:13.0.CO;2-Q

Related works:
This item may be available elsewhere in EconPapers: Search for items with the same title.

Export reference: BibTeX RIS (EndNote, ProCite, RefMan) HTML/Text

Persistent link: https://EconPapers.repec.org/RePEc:bla:jamest:v:41:y:1990:i:1:p:61-66

Ordering information: This journal article can be ordered from
https://doi.org/10.1002/(ISSN)1097-4571

Access Statistics for this article

More articles in Journal of the American Society for Information Science from Association for Information Science & Technology
Bibliographic data for series maintained by Wiley Content Delivery ().