Research on KNN algorithm based on MapReduce in spatial database
Liu Biao of Dalian Maritime University
In this paper, we first try to design a inverted grid index and a spatial KNN query based on MapReduce in cloud environment. The main work of this paper is as follows: (1) aiming at data points in two-dimensional space, this paper designs a distributed inverted grid indexing method, which conforms to the standard of spatial data index. Because the inverted grid index has a loosely coupled and unshared special structure, the index is suitable for parallel queries based on MapReduce for large-scale space-query data. (2) In this paper, a method based on MapReduce for spatial inverted grid indexing is proposed and a parallel KNN query algorithm is mrcircletrip on the basis of the index. In addition, the mathematical proof of the convergence of the algorithm is given in this paper to prove the accuracy of the algorithm's cyclic stop condition. (3) In order to verify the scalability of the indexing structure and the performance of KNN query algorithm, a lot of experiments have been done to establish inverted grid index and KNN spatial query.
Key words: Spatial index KNN query grid index MapReduce
[Download Address]:http://bbs.chinacloud.cn/showtopic-14074.aspx
