Research And Implement On Nearest Neighbor And Reverse Nearest Neighbor Spatial Keyword Query System

With the rapid development of information technology and geographic location technology,there are plenty of text data about the geographic location information on the web,named spatial-textual data.People often use a variety of location-based service applications to take full advantage of these spatial-textual data.Spatial keyword query is one of the techniques commonly used in location-based service applications,which lead to the increasing importance of spatial keyword query.In this paper,we do research on nearest neighbor and reverse nearest neighbor queries in spatial keyword query system,which can be applied to deal with queries from users and merchants respectively.For the different query problems,the corresponding index structures and efficient query processing algorithm are put forward.The contributions of this paper are shown as follows:(1)We propose a mixed index based on G-tree and corresponding query pruning strategies to deal with textual-restricted k nearest neighbor spatial keyword query on road networks.It eliminates the visit of unnecessary of tree nodes by evaluating upper and lower bound between the query object and tree nodes.We have the priority to access the closer nodes of the proposed tree index such that the efficiency of the spatial keyword queries can be improved.(2)We propose an approximate querying algorithm based on MIR-tree to deal with top-m reverse spatial and textual k nearest neighbor query.This algorithm aggregates top-k queries for all users,and approximately finds the best m objects with the continuous iteration instead of computing all combinations.This way can effectively solve the problem of maximizing the result of reverse spatial and textual k nearest neighbor query for multi-object.(3)The experimental results demonstrate the high efficiency of the proposed index structures and query processing methods in this paper by comparing with the existing methods in the real data sets.Besides,we implement the nearest neighbor and reverse nearest neighbor query prototype system.