Authors:
Kiyotaka Takasuka
;
Minoru Terada
and
Kazutaka Maruyama
Affiliation:
The University of Electro-Communications, Japan
Keyword(s):
Web page recommendation, Collaborative filtering, Implicit user profile.
Related
Ontology
Subjects/Areas/Topics:
Artificial Intelligence
;
Knowledge Discovery and Information Retrieval
;
Knowledge-Based Systems
;
Soft Computing
;
Symbolic Systems
;
Web Mining
Abstract:
Because the number of Web pages is very huge, and still increasing, many people have difficulty to reach pages they want. Although social bookmarking and search engines are helpful, users still have to find pages themselves.
Our goal is to recommend Web pages which are supposed to be interesting for a user, without active effort
by the user. We first analyzed the http traffic data in our university collected by a sniffer, and developed a recommendation system that works on URLs and their viewers (IP address). Our system has four features: (1) collaborative filtering, (2) implicit build of user profiles, (3) exclusion of popular Web pages (4) and use of the real activity in our university. We evaluated the effectiveness of our system by applying it to the real http transaction data and found that there were 18 successful recommended users out of 50 users.