无索引空间数据库的基于最优点的集合最近邻查找算法

An Algorithm of Aggregate Nearest Neighbor Query for Non-Index Spatial Database

针对度量空间中的无索引空间数据库,提出一种基于最优点的集合最近邻查找算法及其改进算法.采用真实数据集与人工生成的数据集对算法进行测试,评估所提出算法的效率.实验结果表明,所提算法的效率优于组最近邻居查询算法,并且对于高维数据空间,所提出的算法有较高的稳定性.由于查询区域中数据点的数量比较少,改进的基于最优点的集合最近邻查找算法的效率总体上要比改进前高.

The vantage point-based aggregate nearest neighbor query algorithm and it′s improved algorithm are proposed for non-index spatial database in metric space.The algorithms are tested by using both real datasets and synthetic datasets to evaluate efficiencies of the proposed algorithms.The experimental results show that the efficiencies of proposed algorithms are better than that of multiple query method.Furthermore,the proposed algorithms have higher stabilization in high-dimensional data space.Since there are less data points in it′s search region,efficiency of the improved vantage point-based aggregate nearest neighbor query algorithm is generally higher than the vantage point-based aggregate nearest neighbor query algorithm.