Document Type
Dissertation
Date of Award
Spring 5-31-1998
Degree Name
Doctor of Philosophy in Computing Sciences - (Ph.D.)
Department
Computer and Information Science
First Advisor
Peter A. Ng
Second Advisor
Murat Tanik
Third Advisor
Daochuan Hung
Fourth Advisor
Ronald S. Curtis
Fifth Advisor
Taiming Chu
Abstract
Browsing is a technique, which helps users to formulate their query and retrieve information in the information retrieval system. This technique provides users with capabilities of understanding their information needs and gaining system knowledge during the course of the browsing and thus it eases the users' burden when issuing queries. The basic components of the browser provides an underlying structure which allows users to navigate and a browsing process controller which provides users with the needed assistance during each browsing session.
In this dissertation, a new infrastructure (OP-Net), transformed from the existing object network is proposed. Each object in the object network is transformed into a predicate-augmented information repository. The predicate associated with each information repository governs the content of relevant documents in the depository during the browsing process and is updated continuously according to queries given by the user. The OP-Net with the relevant information repositories provides a dynamic and efficient environment for browsing.
A new ranking model is also proposed based on the signature of the documents and the user's query. The signature of a document is a document representative which utilizes the information provided by the dual model in TEXPROS (TEXt PROcessing System). With the signatures, the similarity of the document and the query can be computed, and the ranks of the documents can be derived.
This dissertation describes a three-layer architecture for the browser. At the top layer, the browsing process controller conducts and monitors the browsing process, and utilizes the services provided by the service providers. At the bottom of this architecture is the storage management system which stores the documents and then associated frame instances and responses to the requests from the service providers in the second layer. This architecture supports the principle of information hiding by allowing the change of the design of each component without changing the others. In the conclusion of this dissertation, the potential improvements and future research will be proposed.
Recommended Citation
Wang, Chih-Ying, "The intelligent browser for texpros" (1998). Dissertations. 959.
https://digitalcommons.njit.edu/dissertations/959
