CAJ | 학술논문

多维数据的处理已经成为影响很多领域发展的关键因素，特别是多维数据的相似性查询已经被用在很多领域中。当数据维度很大的时候，大多数索引结构处理的性能下降，这现象被称为“维度灾难”。针对多维度灾难，RAKDB-Tree是本文提出的一种高效处理多维数据的索引结构。该索引结构首先把数据空间划分为子空间，然后使用改进的KDB-Tree对子空间建立索引。RAKDB-Tree的查询、插入、删除等算法使得，索引结构一直保持较优状态。实验结果表明，RAKDB-Tree能够很好解决因为数据维度增加而带来的各种问题。
다유수거적처리이경성위영향흔다영역발전적관건인소，특별시다유수거적상사성사순이경피용재흔다영역중。당수거유도흔대적시후，대다수색인결구처리적성능하강，저현상피칭위“유도재난”。침대다유도재난，RAKDB-Tree시본문제출적일충고효처리다유수거적색인결구。해색인결구수선파수거공간화분위자공간，연후사용개진적KDB-Tree대자공간건립색인。RAKDB-Tree적사순、삽입、산제등산법사득，색인결구일직보지교우상태。실험결과표명，RAKDB-Tree능구흔호해결인위수거유도증가이대래적각충문제。
For many application areas, the efifciency of multidimensional data processing is a key factor affecting their development. In particular, similarity search is used in many ifelds, such as data mining, big data analysis, digital multimedia etc. However, lots of index structures cannot avoid the“dimensionality curse”, when number of dimensions is very large. RAKDB-Tree uses partitioning method to divide data regions and create approximation ifles. Then RAKDB-Tree indexes the approximations with the improved method of KDB-Tree. RAKDB-Tree is an automatically adjust and optimize tree index structure. Experimental results show that the RAKDB-Tree has a promising improvement in performance.