通过分析已有的索引结构在进行k近邻查询时效率上的不足,提出了适合进行k近邻查询的X*树索引结构,采用了新的结点分裂算法,同时不需要额外存储结点分裂的历史信息。实验结果表明它比X树的时间和空间性能更好,更适合k近邻查询的应用。
By analyzing the inefficiency of k nearest neighbor query in existing index structures,this paper presents X*tree index structure which is suitable to perform k nearest neighbor query.A new node splitting algorithm is adopted,and the Split History field is omitted.The experiment shows that it has better performance than X*tree in time and space complexity,and it is more suitable to k nearest neighbor query.