Dissertation Information for Padmini Das-Gupta
- Padmini Das-Gupta
- Library and Information Science
- Syracuse University (USA) (1985)
- Jeffrey H. Katzer
MPACT Status: Incomplete - Ambiguous
Title: AN INVESTIGATION INTO THE TWO-POISSON MODEL OF AUTOMATIC INDEXING
Abstract: This thesis aims at contributing to the design of document retrieval systems (DRS). A DRS is a system that when given an information need from a user, attempts to retrieve citations to documents which have information of potential relevance to the user. The representation of documents or document indexing is a function basic to a DRS. This thesis examined this function within the framework of the Two-Poisson model of automatic indexing.
The Two-Poisson model is one of the models that have been proposed in the literature for the automatic indexing of documents. In this thesis, the model was examined with two objectives. The first was to improve the model. Herein, specific ways of improving the model, resulting from a theoretical investigation, were tested. In general, the results show that the modifications were only moderately successful in improving the performance of the model. The second objective was to compare the Two-Poisson model to other (statistical) models for automatic indexing. The comparison showed that the Two-Poisson model's performance is similar to those of the other models tested.
This thesis contributes to our understanding of the index term property underlying the Two-Poisson model: the distribution of an index term's within document frequencies.