SOCEN-CLUSTER: A new rapid clustering technique

Abstract

In this paper, we present a novel clustering technique for unindexed, randomized, multidimensional, datasets. The main advantage of the proposed technique is the time and space complexity that were reduced to linear cardinality dependency. The algorithmic implementation shown in this paper uses some heuristics to enhance the overall execution time and space required making them fully scalable. This particularity makes it easier for ASICS / FPGA architects to implement such a technique in a constrained environment.