KNN查询处理算法性能研究

Performance Study of KNN Query Processing Algorithm

KNN查询是多媒体数据库管理系统中最具代表性的查询方式之一,它将k个与查询点最接近的对象作为查询结果返回。对于树型多维索引结构,KNN查询处理算法主要有RKV算法和HS算法。本文针对这两种不同处理算法进行了性能研究,通过试验确定了算法的不同适用场景,最后就应用中的KNN查询实现给出了相应的建议。

KNN query is one of the most representative queries in multimedia database management system. It shows the result of k objects nearest to the query point. There are two different approaches： RKV algorithm and HS algorithm to processing KNN query for multidimensional index structures. In this paper, a performance study of the two different algorithms is presented. Computer simulation results and applicable situations for these two algorithms are also given. Finally some corresponding suggestions are given for applying it.