The paper introduces a new algorithm for a cluster analysis named BClusGe. It uses a technique of the subspace clustering with an empty space removal. This new algorithm produces a multidimensional rectangular grid tightly adjusted to the processed data distribution. Afterwards, BClusGe merges adjacent folds contain-ing objects and produce the output set of clusters. An execution time of BClusGe depends on a number of folds created at the partitioning to describe the data distri-bution. Thus, data with the irregular distribution are clustered slower. Additionally, BClusGe delivers a B-tree structure describing found clusters. Its leafs are interconnected and allow to navigate between dense subspaces of the data distribution. The tree is useful in practical applications because it organizes hierarchically dependencies between clusters. The paper contains some experiments' results where BClusGe performance was compared to DBScan and K-means.
JavaScript jest wyłączony w Twojej przeglądarce internetowej. Włącz go, a następnie odśwież stronę, aby móc w pełni z niej korzystać.