A tag recommendation system for folksonomy

  • Authors:

  • Ning Zhang;Yuan Zhang;Jie Tang

  • Affiliations:

  • Tsinghua University, Beijing, China;Tsinghua University, Beijing, China;Tsinghua University, Beijing, China

  • Venue:
  • Proceedings of the 2nd ACM workshop on Social web search and mining
  • Year:
  • 2009

Quantified Score

Hi-index 0.00

Visualization

Abstract

Social bookmarking tools become more and more popular nowadays and tagging is used to organize information and allow users to recall or search the resources. Users need to type the tags whenever they post a resource, so that a good tag recommendation system can ease the process of finding some useful and relevant keywords for users. Researchers have made lots of relevant work for tag recommendation systems, but those traditional collaborative systems may not perform well in the case of a new user and a new resource. To address this problem, we propose a tag recommendation system for two different cases. The first case is that the active user or active resource is a new one which has not appeared in the past while the second case is that the active user and resource has already been posted. In this paper, we present two different methods for these two cases: a content-based method and a graph-based method. We implement our tag recommendation system in a realworld academic Folksonomy system. In order to gain better results,we apply improvements and combinations of previous work in each of these two methods.and combinations of previous work in each of these two methods. We participated into the Discovery Challenge 2009 with the proposed methods. The results of our teams are within the top teams.