Overviewing the Knowledge of a Query Keyword by Clustering Viewpoints of Web Search Information Needs

In this paper, we address the issue of how to overview the knowledge of a given query keyword. We especially focus on concerns of those who search for Web pages with a given query keyword, and study how to efficiently overview the whole list of Web search information needs of a given query keyword. First, we collect Web search information needs of a given query keyword through search engine suggests. In the case of a Japanese search engine, we collect up to around 1,000 suggests given a query keyword. However, some of them are redundant in that they originate from almost the same Web search information needs. In order to aggregate such redundant search engine suggests, we take an approach of clustering search engine suggests based on document vectors generated from the snippets shown by the search engine. Evaluation result shows that the proposed clustering approach proves to be quite useful for efficiently over viewing Web search information needs of a given query keyword. We also develop an interface system for over viewing those aggregated search engine suggests of a given query keyword as well slinks to top ranked Web pages that are closely related to those aggregated search engine suggests.