A thesis submitted in partial fulfillment of the requirements for the degree of Master of Science, School of Advanced Technologies

Thesis (M.Sc.) - Asian Institute of Technology, 1998

The promise of the information age entails controlling the potentially unwanted flux of information from sources to their receiving end-users and deliver just the relevant information. The servers on the Web represent a vast potential resource. As the number of these Internet servers increases rapidly, it becomes difficult to determine the relevant servers when searching for information. To address this issue, it is necessary to formalize the querries and the servers and develop a measure of similarity between them. We have reviewed the literature available in this area, paying attention to similarity measure. We propose a new similarity measure to rank information filtering system for Boolean queries. The focus is on Boolean environments as it is generally believed that Boolean expressions can precisely describe a server's contents as well as a user's information need. Users can send queries to the top-ranked systems and obtain most of the relevant information. To compare the performance of the existing measure, we conducted two case studies on compendex databases. The thesis includes the details of the case study experiments and performance comparisons of the proposed measure with those available. The results are found to be encouraging.