Abstract | ||
---|---|---|
The evolution of the WWW has led to an explosion of information and consequentially a significant increase on usage. This avalanche effect has resulted in such uncertain environment in which we find it difficult to clarify what we want, or to find what we need. In this paper we introduce RecSys which aims to confront the problem by developing a software agent which intelligently learns users interests, and hence makes recommendations of resources on WWW based on the user's profile. The system employs multiple TFIDF vectors to represent various domains of user's interests. It continuously and progressively learns users profile from both implicit and explicit feedback. This is achieved by extraction and refinement of featured keywords within the learning examples. Several heuristics were also adapted to improve the overall performance of the system. |
Year | Venue | Keywords |
---|---|---|
2008 | ICEIS 2008 : PROCEEDINGS OF THE TENTH INTERNATIONAL CONFERENCE ON ENTERPRISE INFORMATION SYSTEMS, VOL SAIC: SOFTWARE AGENTS AND INTERNET COMPUTING | Recommender systems,web browsing,user modelling |
Field | DocType | Citations |
Data mining,World Wide Web,Software engineering,Computer science,Software agent | Conference | 0 |
PageRank | References | Authors |
0.34 | 1 | 1 |
Name | Order | Citations | PageRank |
---|---|---|---|
Gulden Uchyigit | 1 | 5 | 5.53 |