Presenting Query Aspects to Support Exploratory Search

Successful information search requires a joint effort from both syntactic matching provided by current search engines and semantic matching performed by human users. Word-based syntactic matching schemes work well for tasks such as homepage finding or fact finding, but they are less effective in supporting exploratory search tasks such as learning and investigation. One way to overcome this limitation of syntactic matching is to capture the search journeys of other users with semantically related queries, and use them as a roadmap to guide exploratory search. This paper presents our investigation on the utilization of query semantics derived from query logs, to 1) increase the diversity of a search result; and 2) devise new interfaces that display a search result to support exploratory search. We conducted a user study to evaluate our initial interface prototypes. The evaluation shows that, with the interface that explicitly supports their task, subjects acquire more knowledge and are more confident about their task completeness. The differences between subjects’ preferences suggest that we may need to provide a range of interfaces that can not only support users’ search tasks, but also suit their personal styles.