Discovery of shared topics networks among people: a simple approach to find community knowledge from WWW bookmarks

  • Authors:
  • Hideaki Takeda;Takeshi Matsuzuka;Yuichiro Taniguchi

  • Affiliations:
  • National Institute of Informatics, Chiyoda-ku, Tokyo, Japan and Graduate School of Information Science, Nara Institute of Science and Technology, Ikoma, Nara, Japan;Toppan Printing Co.,Ltd. and Graduate School of Information Science, Nara Institute of Science and Technology, Ikoma, Nara, Japan;Graduate School of Information Science, Nara Institute of Science and Technology, Ikoma, Nara, Japan

  • Venue:
  • PRICAI'00 Proceedings of the 6th Pacific Rim international conference on Artificial intelligence
  • Year:
  • 2000

Quantified Score

Hi-index 0.00

Visualization

Abstract

In this paper, we propose a system called kMedia that can assist users to form knowledge for community by showing shared topics networks (STN) among them. One of the important aspects to know each other is to know topics interested by others and relationship between her/his and others' topics. kMedia can use a simple but effective way to find them. It uses folders in WWW bookmarks as interested topics and can calculate their relations by evaluating similarity of WWW pages under folders. The results are displayed in two ways. One is to show relationship among users by shared topics networks, i.e., a user is connected to the other through both her/his topics and the other's topics that are related to her/his ones. A user can know what kind of relations to others s/he can have, and more precisely know what are counterpart of her/his topics for others. The other way is to show recommended pages for pages in users' bookmarks. Recommended pages are selected from others' bookmarks, and it is the primary result of similarity evaluation among pages by contents. A user can use this result just as recommendation for her/his bookmarked pages or use checking how her/his bookmarked pages are related to others. We tested this system in an experiment with actual bookmark data. Discovery of related topics among users are evaluated as good enough in spite of bad results for recommendation of pages. This result tells that our approach to find common topics among users is effective and practical.