Research And Implementation Of Recommendation System Based On Reverse K Nearest Neighbor Query In LBSN Environment

With the advent of the age of big data,the location-based social network service platform has developed rapidly.The recommendation technology based on the reverse k nearest neighbor query plays an important role in the service of the LBSN platform.LBSN data has multiple heterogeneity,real-time update and spatial text.Therefore,compared with traditional reverse k nearest neighbor query in LBSN environment,there are problems such as large data scale and frequent data update.In view of the above problems,this paper proposes a recommendation system based on the reverse k nearest neighbor query in the LBSN environment.Aiming at the problem of continuous bicolor reverse k nearest neighbor query in a data stream environment,this paper first proposes an MHTree index.Its purpose is to efficiently maintain the result set when the data is updated.Based on this,this paper proposes a Six-Region ring query processing algorithm to achieve high-quality data filtering.Finally,this paper proposes a data preprocessing method to efficiently maintain the candidate user set and verification facility set,which improves the performance of algorithm update.To solve the problem of distance-keyword similarity bicolor reverse k nearest neighbor query,this paper proposes a filter-validation query processing framework based on the Keyword Multiresolution Grid Rectangle-tree.Firstly,the framework proposes a KMG-Tree index to achieve efficient management of spatial text database.Secondly,this paper proposes a SixRegion-Optimize algorithm to achieve efficient pruning of the KMG-Tree index.Finally,this paper uses the threshold ?r to screen out a small number of objects of higher quality.Finally,the SRR and Six-Region-Optimize algorithm are verified by using part of the geospatial position information from Germany and North and a simulated data set.The validity,practicability and stability of the proposed algorithm are proved.This paper implements a recommendation system based on reverse k nearest neighbor query in LBSN environment.The system supports two recommendation methods of continuous bicolor reverse k nearest neighbor query under streaming data environment and distance-keyword similarity constraint bicolor reverse k nearest neighbor query.