Information Filtering Using User‘s Context on Browsing in Hypertext

  • Authors:
  • T. Hirashima;K. Hachiya;A. Kashihara;J. Toyoda

  • Affiliations:
  • Dept. of Artificial Intelligence, Kyushu Institute of Technology, 680-4, Kawazu, Iizuka, Fukuoka 820, Japan. e-mail: tsukasa@ai.kyutech.ac.jp;ISIR, Osaka University, 8-1, Mihogaoka, Ibaraki, 567 Japan;ISIR, Osaka University, 8-1, Mihogaoka, Ibaraki, 567 Japan;ISIR, Osaka University, 8-1, Mihogaoka, Ibaraki, 567 Japan

  • Venue:
  • User Modeling and User-Adapted Interaction
  • Year:
  • 1997

Quantified Score

Hi-index 0.00

Visualization

Abstract

Browsing is one of the most popular ways togather information in database with hypertextstructure. In order to support a user to browse,modeling of the user‘s interests is one of the mostimportant issues. Although there are several promisingmethods to infer the interests from the user‘sbrowsing behavior, they assume that the interests areconsistent during the browsing. However, the user‘sinterests are often strongly dependent on the localcontext of the browsing. This paper describes a methodto model the user‘s shifting interests from thebrowsing history. An information filtering methodusing the model of the interests has been implemented.We call it ’context-sensitive filtering‘. The resultsof an experimental evaluation, by real users‘ browsingfor an encyclopedia in CD-ROM format, are alsoreported.