Selecting Web Search Results of Diverse Contents with Search Engine Suggests and a Topic Model.

In this paper, we address the issue of how to overview the knowledge ofa given query keyword. We especially focus on concerns of those whosearch for Web pages with a given query keyword, and study how toefficiently overview the whole list of Web search information needs of agiven query keyword. First, we collect Web search information needs ofa given query keyword through search engine suggests. Although wecollect up to around 1,000 suggests given a query keyword, some of themare redundant in that they originate from almost the same Web searchinformation needs. In order to aggregate such redundant search enginesuggests, we take an approach of clustering search engine suggests basedon a topic model. Evaluation result shows that the proposed clusteringapproach proves to be quite useful for efficiently overviewing Websearch information needs of a given query keyword. We also develop aninterface system for overviewing those aggregated search engine suggestsof a given query keyword as well as links to top ranked Web pages thatare closely related to those aggregated search engine suggests. Finally, we show the effectiveness of the interface in terms of theaggregation of Web search results.