| Due to the past decade advances in the computer technologies, the Internet becomes the largest, ultimate and most used source of information. The Internet now seems to have an inexhaustible supply of information on every conceivable subject. To many, this is its major fascination, attraction, and strength. However, the opposite case is also true, in that the quantity of information available is a weakness. The sheer volume of information to search is almost unmanageable. The results returned from searches by conventional search engines often supply information that cannot possibly be analyzed in a reasonable time frame. For example, a search using a conventional search engine such Infoseek asked this simple, straightforward question: "Can you find me a job?" The search engine returned 39,746,521 pages. If we allow 30 seconds to review each page, it would take 37 years and 6 months, with no breaks, to review all the pages. This review would be a full-time occupation in itself! Often we are overwhelmed by these search results and so abandon them. If we look hard and long enough, we might find what we want, or at best, we compromise our search or the results we are seeking.; Thus, users need an intelligent and efficient information search tool that could relieve them from the hard task of reviewing a large amount of results. This thesis proposes an agent, named ARIII, which fulfills this purpose. The main function of this agent is to discover and retrieve on the behalf of the user pertinent information over the Internet by taking in consideration the user profile and performing a pertinent sort and filtering of the results that matches the user's preferences and interests. |
