【摘要】 提出了一种新的M+-tree构造策略—通过批量装载数据静态构造M+-tree.这种建树方法能够极大地降低索引的构建代价并加快建树速度,又能保证高维数据良好的聚类特性.同时又提出了一种有效的聚类算法—多步聚类,基于这种聚类构造M+-tree.实验结果表明,这种建树算法能够保证索引空间的最大利用率,降低建树的代价,同时又能够保证索引的查询效率.更多还原

【Abstract】 This paper proposed a novel strategy for constructing M+-tree—static bulk—loading M+-tree. This method had a good data clustering while greatly reducing the cost of index construction and speeding up this process. At the same time, we proposed an effective multi-steps clustering algorithm, based on which,M+-tree is well built.Our experiments showed that this constructing method could ensure the maximal space utilization,reduced the cost of index construction,and improve the query performance of index. [HJ]更多还原