Research On Spatial Text Query Based On Geohash

Spatio-textual query is widely used in location-based service(LBS),which has received extensive attention from researchers in recent years.As an important research topic in the field of spatio-textual processing,the essential of spatio-textual query is a cross-fusion of spatial query and textual query.At present,the research work of spatio-textual query mostly focuses on designing efficient spatial index,textual index and combination way.With the increase of spatial data size and dimension,the traditional multi-dimensional indices becom more and more complex to increase the time cost of querying and reduce the efficiency of spatio-textual query.Since the space filling curve can convert the problem of spatial range query into string prefix matching problem,Trie can efficiently supports prefix search by compressing the common prefix of string into a path,therefore,this paper constructs a spatial index structure named GeoExTrie which combines Geohash and Trie.To effectively support spatial range query,Trie node is extended by adding the begin ID of the current node prefix and its number,which allows only access the path from root to internal node without to leaf node for a wide range of spatial query.Based on GeoExTrie,this paper designs a filtering-verification framework to implement spatial range query.In filtering phase,obtaining candidate set by retrieving GeoExTrie.In verification phase,this paper proposed a method of verifying according to the range of latitude and longitude range to return location that satisfies query-condition.The experimental show that the efficiency of searching GeoExTrie better than R-Tree and R*-Tree when range query less than 1000 m.At present,most spatio-textual indices focus on the combination of R-Tree or its basic variants and inverted list to achieve spatio-textual querying in the field of spatio-textual query,and less research work is used to explore more spatio-textual indices;and as the amount of spatio-textual data increases,the complexity of multi-dimensional index structure based on R-tree or its basic variants directly leads to reduce the efficiency of spatio-textual query.In order to solve this problem,TextGeoExTrie is proposed to process spatial range textual subset query,combinedGeoExTrie and inverted list,which is compact and highly efficient.In order to improve the efficiency of textual retrieval,the data is sorted by spatial-filling-curve order,so that only a small part of inverted list needs to be accessed when searching avoid scanning the entire list,which improves query efficiency.This paper compares TextGeoExTrie with SFC-Quad and KR*-Tree,the classical indices in the domain of spatial range textual subset query.The experimental results show that the response time of SFC-Quad or KR*-Tree query increases linearly,and TextGeoExTrie has higher retrieval efficiency with the increasing amount of spatio-textual data.In summary,the main work of this paper is to realize arbitrary spatial range query by combining Geohash and Trie,and then combine the inverted index to realize spatial range textual subset querying,and point out the next work plan.The experimental results show that the combination of Trie to process spatial query or spatio-textual query is effective.