An approach for faster processing of top-k spatial keyword queries

Abstract

A typical spatial keyword query (SK-query) takes a user location and a set of keywords as arguments and returns objects that are spatially and textually relevant to the query. The top-k SK-query retrieves top-k results ranked according to a score that takes into consideration spatial proximity and textual relevance. Index structures like RTree and inverted indexes are used for the processing of all SK-queries. Hybrid index structures like HybridI, HybridR and also IR-tree index combine these access methods. Approach based on IR-tree provides efficient document ranking, spatial and textual filtering but do not take into account the unnecessary disk accesses it incurs. In this paper, we propose a method to reduce unnecessary disk accesses by employing a storage scheme that enhances the existing processing of SK-queries using IR-tree index. This can improve the processing of top-k spatial keyword queries by speeding up the scoring or ranking of the documents.