SOaP: Social Filtering through Social Agents

The Web is becoming the premium source of information for a growing number of people. As a result, information overload arises as a problem of extracting useful information. Information gathering on the Web has become a time-consuming work. As an emerging technique for dealing with this problem, collaborative filtering (also known as social filtering) can improve retrieval precision and reduce the amount of time spent. In this paper we propose a social filtering system consisting of various types of agents which mediate between different people, groups and the Web. Agents work on behalf of their clients users or other agents under the specified security and/or privacy constraints. They interact with each other and allow people to cluster the URLs, rate and annotate the Web pages, and share the recommendations. Agents could also find people and groups with similar interests, bring people to- gether to form groups and allow them to work within various groups to exploit the collected bookmarks. Eventu- ally, the system could contribute to the social construction of knowledge on the Web.