Multiple Domain-Based Spatial Keyword Query Processing Method Using Collaboration of Multiple IR-Trees

Abstract

In this paper, we propose multiple domain-based spatial keyword query, called the MDR (Multiple-Domain based Range) query, and query processing algorithm. The MDR query consists of two sub-queries. The first sub-query, called the primary range query, identifies a group of geo-textual objects that are within a query range. The second sub-query, called the refining range query, identifies the geo-textual objects whose neighboring geo-textual objects contain the desired keywords from the results of the primary range query. The problem is that the keywords domains of each sub-queries are different. Existing methods for spatial keyword queries cannot handle the MDR query efficiently because they are designed to handle the keywords of a single domain. In order to accommodate the keywords of various kinds of domains concurrently, we divide the geo-textual database by the domain of the keyword and construct IR-trees for each domain. Based on this approach, our query processing algorithm traverses multiple IR-trees simultaneously, in order to reduce the number of the refining query. Our performance studies show that our searching strategy significantly reduces the query response time.