Abstract
The rapid growth of web pages makes searching and retrievinginformation on web efficiently and effectively a criticalproblem. While common keyword-based search engines bringvague and numerous results, we introduce the concept toquery against. URLs which would bound the search results into amuch smaller fraction. As conventional informationfiltering services, in which users submit profiles such thatthey will be automatically informed of new addtions that maybe of intrest, we expect URL profiles will soon be applied onsimilar applications.In this thesis we propose several index structures for indexingprofiles and algorithms that efficiently match URLs against alarge number of profiles. We also present simulation resultsto compare their performance under different scenarios.