Dynamic term suggestion for searching multilingual school documents

  • Authors:

  • Kohei Sawa;Yusuke Okano;Masahiro Hori;Chigusa Kita

  • Affiliations:

  • Graduate School of Informatics, Kansai University, Takatsuki-shi, Osaka, Japan;Faculty of Informatics, Kansai University, Takatsuki-shi, Osaka, Japan;Graduate School of Informatics, Kansai University, Takatsuki-shi, Osaka, Japan;Graduate School of Informatics, Kansai University, Takatsuki-shi, Osaka, Japan

  • Venue:
  • Culture and computing
  • Year:
  • 2010

Quantified Score

Hi-index 0.00

Visualization

Abstract

The number of school children having ties overseas is on the rise year after year in Japan. In order to support these children and their parents, we developed a multilingual school document portal site, and made it open to the public. The portal site allows easy-to-use document retrieval by faceted classification as well as keyword search. However, it is not necessarily easy for the users to express their information needs in query terms, and they often come up with poor results from short and generic query terms. In pursuit of the formulation of better query statements with a few search terms even in the initial query, we realized a dynamic term suggestion or auto-suggest interface in the portal site. In the auto-suggest interface, suggested terms are ranked according to domain relevance, rather than merely on the basis of term occurrence frequency in the document collection. In this paper, we explain an overview of the school document portal site, the auto-suggest interface realized in the portal, and the domain-dependent term-weighting scheme, followed by the results of a user study conducted to demonstrate the effectiveness of the auto-suggest function.