Suggesting Sub-Topics of Issued Queries by Using Concept Structure
Kazuki Hagiwara, Kenji Hatano · Frontiers in artificial intelligence and applications · 2016
In this paper, we propose a method for suggesting sub-topics of an issued queries by using the conceptual structure of WordNet. In existing search engines, sub-topics are created on the basis of the input histories of search engine users. Therefore, often-used pairs of key- and sub-topics can be issued; otherwise it is difficult for unaccustomed users to input both key- and sub-topics. We believe that the proposed method can help the users to issue appropriate search queries conveying their own information needs because sub-topics are general subsidiary information of a key-topic and are automatically suggested as the context of a key-topic irrespective of the input histories. To evaluate the proposed method, we conducted a usability test of the created search queries using a crowdsourcing site and compared the usability of the proposed method with that of the existing search engines.