No abstract

Recently data mining applications require a large an~uut of highdimensional data, However, most clustering methods for data miming do not work efficiently for dealing with large, highdimensional dam because of the so-called 'curse of dimensionalky'[l] and the limitation of available memory. In this paper, we propose a new cell-based clustering method which is more efficient fur large, high-dimensional data than_ the existing clustering methods. Our clustering method provides an efficient cell creation algorithm using a space-partitioning technique and uses a filtering-based index slmcture using an approximation technique. Finally, we compare the performance of onr cell-based clustering method with the CLIQUE mothod in ~ms of cluster construction time, precision, and rclfieval tim=. The experkncntal results show that our clustering method achieves better performance on cluster couslxuctiou time and retrieval time. KeywordsCell-based clustering, filtering-based index dimensional data, data mining. s~eum=, high

No abstract

No abstract

No abstract