The Effectiveness of Latent Semantic Analysis for Building Up a Bottom-up Taxonomy from Folksonomy Tags

  • Authors:

  • Takeharu Eda;Masatoshi Yoshikawa;Toshio Uchiyama;Tadasu Uchiyama

  • Affiliations:

  • NTT Cyber Solution Laboratories, NTT Corporation, Kanagawa, Japan;Kyoto University, Kyoto, Japan;NTT Cyber Solution Laboratories, NTT Corporation, Kanagawa, Japan;NTT Cyber Solution Laboratories, NTT Corporation, Kanagawa, Japan

  • Venue:
  • World Wide Web
  • Year:
  • 2009

Quantified Score

Hi-index 0.00

Visualization

Abstract

In this paper, we evaluate the effectiveness of a semantic smoothing technique to organize folksonomy tags. Folksonomy tags have no explicit relations and vary because they form uncontrolled vocabulary. We discriminates so-called subjective tags like "cool" and "fun" from folksonomy tags without any extra knowledge other than folksonomy triples and use the level of tag generalization to form the objective tags into a hierarchy. We verify that entropy of folksonomy tags is an effective measure for discriminating subjective folksonomy tags. Our hierarchical tag allocation method guarantees the number of children nodes and increases the number of available paths to a target node compared to an existing tree allocation method for folksonomy tags.