Christopher H. Brooks, Yeh Fang, Ketaki Joshi, Xia Zhou, Papanii Okai
This paper describes CitePack, an autonomous, agent-based system for automatically discovering resources based on learned user preferences. We discuss the need for a tool that can find information for a user without direct intervention, followed by the architecture of CitePack. We discuss the collection of related documents, and the construction of a user model based up support vector machines, which allows us to both learn a user's tastes and also recommend similar users. We conclude by discussing potential future directions.
Subjects: 1. Applications; 1.10 Information Retrieval
Submitted: May 14, 2007