Efficient Processing of Top-k Spatial Keyword Queries

Abstract

Abstract. Given a spatial location and a set of keywords, a top-k spatial keyword query returns the k best spatio-textual objects ranked according to their proximity to the query location and relevance to the query keywords. There are many applications handling huge amounts of geotagged data, such as Twitter and Flickr, that can benefit from this query. Unfortunately, the state-of-the-art approaches require non-negligible processing cost that incurs in long response time. In this paper, we propose a novel index to improve the performance of top-k spatial keyword queries named Spatial Inverted Index (S2I). Our index maps each distinct term to a set of objects containing the term. The objects are stored differently according to the document frequency of the term and can be retrieved efficiently in decreasing order of keyword relevance and spatial proximity. Moreover, we present algorithms that exploit S2I to process top-k spatial keyword queries efficiently. Finally, we show through extensive experiments that our approach outperforms the state-of-the-art approaches in terms of update and query cost.

ST-HBase: A Scalable Data Management System for Massive Geo-tagged Objects

“…Rocha-Junior et al [1] proposed another novel method that can process topk spatial keyword query more efficiently. Differently from [3], Rocha-Junior et al proposed a new index structure called Spatial Inverted Index (S2I).…”

“…They adopted different organization methods for the terms based on their frequency. It is well known that the distribution of terms is very skewed, the document frequency of terms in a corpus follows the Zipf's law, which means that only a small number of terms occur frequently, while most of the terms occur infrequently [1] [4]. So S2I mapped each more frequent term to a distinct aggregated R-tree (aR-tree) that stores the objects with the given term, if the number of the objects corresponding to one given term does not exceed a given threshold, the objects are just stored in a file.…”

“…In this section, we compare our approaches with the Spatial Inverted Index(S2I) proposed by Joao et al [1], to the best of our knowledge, S2I is the state-of-theart approach that addresses spatial keyword queries. The setup for S2I is the same as described in [1], the nodes of the aR-trees have a block size of 4KB and are able to store between 42 and 85 entries.…”

“…Figure 1-(a) shows the performance of R-tree index, it just uses R-tree to index the spatial information but neglects the textual information. shows that the performance of the Hybrid index S2I [1]. S2I [1] took both spatial and textual information into consideration and proposed a hybrid index.…”

“…shows that the performance of the Hybrid index S2I [1]. S2I [1] took both spatial and textual information into consideration and proposed a hybrid index. From Figure 1, we can observe that the insert throughput is very slow and decreases with the data size increasing, because the R-tree like index needs additional cost to adjust the R-tree node and the inverted index file; the response time increases with data size increasing, because the overlap between the R-tree nodes increases as the data size becomes bigger and bigger.…”