An Index Structure for Efficiently Handling Dynamic User Preferences and Multidimensional Data

R-tree is index structure which is frequently used for handling spatial data. However, if the number of dimensions increases, or if only partial dimensions are used for searching the certain data according to user preference, the time for indexing is greatly increased and the efficiency of the generated R-tree is greatly reduced. Hence, it is not suitable for the multidimensional data, where dimensions are continuously increasing. In this paper, we propose a multidimensional hash index, a new multidimensional index structure Received (May 10, 2017), Review Result (May 24, 2017) Accepted (May 31, 2017), Published (July 31, 2017) 28644 Dept. Computer Science, Chungbuk National Univ., Seowon-gu, Cheongju, Chungbuk, Korea email: leopard@chungbuk.ac.kr 28644 Dept. Computer Science, Chungbuk National Univ., Seowon-gu, Cheongju, Chungbuk, Korea email: khyoo@chungbuk.ac.kr (Corresponding Author) 28644 Dept. Computer Science, Chungbuk National Univ., Seowon-gu, Cheongju, Chungbuk, Korea email: aziz@chungbuk.ac.kr * 본 연구는 산업통상자원부(사업번호: 1005-1028) 지원사업의 연구결과로 수행되었음. An Index Structure for Efficiently Handling Dynamic User Preferences and Multidimensional Data Copyright c 2017 HSST 926 based on a hash index. The multidimensional hash index classifies data into buckets of euclidean space through a hash function, and then, when an actual search is requested, generates a hash search tree for effective searching. The generated hash search tree is able to handle user preferences in selected dimensional space. Experimental results show that the proposed method has better indexing performance than R-tree, while maintaining the similar search performance.