본문 바로가기
  • Home

Mining Semantically Similar Tags from Delicious

  • Journal of the Korean Society for Information Management
  • Abbr : JKOSIM
  • 2009, 26(2), pp.127~147
  • DOI : 10.3743/KOSIM.2009.26.2.127
  • Publisher : 한국정보관리학회
  • Research Area : Interdisciplinary Studies > Library and Information Science
  • Received : May 16, 2009
  • Accepted : June 2, 2009
  • Published : June 30, 2009

Kwan Yi 1

1Univ. of Kentucky

Accredited

ABSTRACT

The synonym issue is an inherent barrier in human-computer communication, and it is more challenging in a Web 2.0 application, especially in social tagging applications. In an effort to resolve the issue, the goal of this study is to test the feasibility of a Web 2.0 application as a potential source for synonyms. This study investigates a way of identifying similar tags from a popular collaborative tagging application, Delicious. Specifically, we propose an algorithm (FolkSim) for measuring the similarity of social tags from Delicious. We compared FolkSim to a cosine-based similarity method and observed that the top-ranked tags on the similar list generated by FolkSim tend to be among the best possible similar tags in given choices. Also, the lists appear to be relatively better than the ones created by CosSim. We also observed that tag folksonomy and similar list resemble each other to a certain degree so that it possibly serves as an alternative outcome, especially in case the FolkSim-based list is unavailable or infeasible.

Citation status

* References for papers published after 2023 are currently being built.