Using Web Snippets and Query-logs to Measure Implicit Temporal Intents in Queries

Understanding the user's temporal intent by means of query formulation is a particular hard task that can become even more difficult if the user is not clear in his purpose. For example, a user who issues the query Lady Gaga may wish to find the official web site of this popular singer or other information such as informative or even rumor texts. But, he may also wish to explore biographic data, temporal information on discography release and expected tour dates. Finding this information, however, may prove to be particularly difficult, if the user does not specify the query in terms of temporal intent. Thus, having access to this data, will allow search mechanisms to improve search results especially for time-implicit queries. In this paper, we study different approaches to automatically determine the temporal nature of queries. On the one hand, we exploit web snippets, a content-related resource. On the other hand, we exploit Google and Yahoo! completion engines, which provide query-log resources. From these resources, we propose different measures to understand the temporal nature of queries. We compare these measures by analyzing their correlation. Finally, we conduct a user study to temporally label queries.