Answering recreational web searches with relevant things to do results

Highlights：

• We propose the problem of recreational queries in information retrieval and propose a solution that combines search query logs with LBSNs.

• We describe a taxonomy for recreational queries derived from real world data by examining a real-world query log (Bing.com)

• We introduce a relevance model that incorporates spatial, temporal information, and social data.

• We detail the offline evaluation of the POIs proposed by our techniques. This topic is usually not covered nor described in detailed in related work. We showed that assessing POIs is very complex and propose design alternatives that work well.

• To summarize, we present an end-to-end data driven system that uses LBSN data to solve a common web search scenario.

摘要

•We propose the problem of recreational queries in information retrieval and propose a solution that combines search query logs with LBSNs.•We describe a taxonomy for recreational queries derived from real world data by examining a real-world query log (Bing.com)•We introduce a relevance model that incorporates spatial, temporal information, and social data.•We detail the offline evaluation of the POIs proposed by our techniques. This topic is usually not covered nor described in detailed in related work. We showed that assessing POIs is very complex and propose design alternatives that work well.•To summarize, we present an end-to-end data driven system that uses LBSN data to solve a common web search scenario.

论文评审过程：Received 27 May 2019, Revised 28 November 2019, Accepted 15 December 2019, Available online 28 December 2019, Version of Record 28 December 2019.