Spatial Keyword Queries With Query Understanding

With the widespread use of smart mobile devices and the rapid development of Mobile Internet in recent years,the Location Based Services(LBS)is playing an increasingly im-portant role in people's daily life.Various LBS-based applications,like Meituan and Baidu Map,have brought a great convenience to people and also have made tremendous impact on our economic,social and cultural life.Spatial Keyword Query(SKQ),one of the most important techniques in LBS,plays a vital role in searching among huge volumes of spatial-textual data.Among all problems in SKQ scenario,the ability of query intention inference fundamentally determines the value of SKQ systems.Existing SKQ techniques mainly infer the query intention by exact/approxiamte matching of query keyword and semantic under-standing.While in the scenario with short-text query,the effectiveness is limited due to the information insufficiency and semantic ambiguity of short text.To address the above difficulties,we study the problem of Spatial Keyword Query with Query Understanding in this paper.First,we propose the interactive spatial keyword query with semantics.It integrates interactive mechanisms with semantic query and can infer the query semantics through user's feedbacks.With the development of knowledge graph,its knowledge-enhanced logistic structure opens the possibilities to short-text query intention inference.On this basis,we propose the conceptual inference based spatial keyword query.It can retrieve a set of conceptual labels by conceptual inference techniques,which reflects the query intention.Besides,with the interactive framework for understanding the distance measure,the user's individual requirements for spatial-textual objects can be well satisfied.Finally,empirical study on real datasets demonstrates the effectiveness and efficiency of our proposed frameworks for query understanding.