DiSCl: Distributed Intelligent Subspace Clustering, a density based clustering approach for very high dimensional distributed dataset

摘要

In this paper, a problem called, “Distributed Subspace Clustering for high dimensional distributed database, based on density notion of clustering” is explored. To solve this problem, we described our algorithm ISC (Intelligent Subspace Clustering) which uses the concept called Hierarchical Subspace Clustering. ISC finds the input parameter ∊ i.e. the distance, required for any density based clustering, adaptively at various levels of dimensionalities. This gives the ability of incremental learning and dynamic inclusion and exclusions of subspaces which lead to better cluster formation.