A Multi-Agent Based Personalized Search Engine with Topical Crawling Capabilities
Disha Verma, Kanika Minocha, Barjesh Kochar · SSRN Electronic Journal · 2014
The tremendous growth in web content and user preferences resulted in a single keyword producing millions of results. Conventional search engines aim at fetching maximum results which match the specified keyword. Our research proposes a search engine which produces fewer but personalized results. The proposed search engine has a layered architecture (multiple agents) which personalizes the results on the basis of different parameters and domain. The domain under consideration is education. The search engine is an amalgamation of client side and server side personalization. The search engine would personalize results on the basis of user browsing history and explicit profile created by him. The server side personalization would work on the profile created by him, whereas user’s browsing pattern would be stored in his personal computer in the form of cookies. For maintaining speed and reliability, the technology used in computation is in-memory data grid.