Research On Reverse Nearest Neighbor Query Algorithm For Spatial Text Objects

With the development of society,Location-based Service has been widely used,resulting in a large number of complex spatial-textual data.However,the data can not be effectively solved by traditional database query techniques.So how to provide a variety of efficient query approaches for spatial-textual objects is one of the hot topics in the field of spatial-textual database.Many kinds of query techniques have been proposed in this field,including Nearest Neighbor(NN)query,Reverse Nearest Neighbor(RNN)query,continuous query and dynamic query and so on.The RNN queries are mostly used for market forecasting and decision support systems,etc.Motivated by this problem,some work has been done as follows in this paper.Firstly,the work of the location selection queries of shops was studied.These queries expect to find a location and service,making the new shop at this location the best choice for as many users as possible.Based on this scenario,a MaxBRAkNN query technique for static,single-label spatial-textual objects was proposed.This type of query aims to find an optimal location and text label,so that the result of BRA:NN query issued by the query object is maximized.Textual objects were classified using the text clustering techniques.The Minimal Bounding Rectangles(MBRs)were constructed based on the extended R-tree.The overall performance is improved by using pruning techniques based on the MBR nodes.Secondly,a new type of MaxBRkNN query for the extended shop location selection decision problem was proposed.According to the nature of the Voronoi diagram,the approximate top-k spatial region of the query object was constructed.Then the spatial region was decomposed based on the quad-tree index structure.The performance is improved,according to the dominance relationship between grids.Further,the work of the decision-making problem for moving objects was studied.Due to the dynamic changes of query objects and user objects,the traditional database query techniques can not perform well.Based on the research of the static MaxBRkNN queries and MOST model in this paper,a DTLUS strategy was proposed,and the decision-making problem for moving objects was solved based on the static MaxBRkNN query algorithms.The MaxBRkNN queries for spatial-textual objects are studied separately under static and dynamic conditions.For the algorithms and techniques proposed in this paper,it is proved in the relevant chapters by experiments that they have a good effect and can be effectively used in the field of decision support.