Efficiently Evaluating Range-Constrained Spatial Keyword Query on Road Networks

With the rapid development of geo-positioning technologies, spatial information retrieval plays an important role in a wide spectrum of applications, e.g., online maps and location-based services. Specifically, spatial keyword query (SK query), considering both spatial proximity to the query location and textual relevance to the query keywords, is now a hot research topic in database community. This paper addresses a specific type of SK query, termed range constrained spatial keyword query (RCSK query), which searches for all the POIs (points of interest) whose textual description is relevant to the query keywords within a specified area. Though RC-SK query has received extensive studies in Euclidean space, little is done to deal with it on road networks. In this paper, alternative approaches with different indexing strategies are proposed to solve this problem. Extensive empirical studies on multiple real datasets demonstrate the efficiency of these proposed approaches.