Multilevel clustering on very large scale of web data

Abstract

With the evolution of the WWW, the computer world has become a huge wave of data, to perform a search of this data, the classical approaches of data mining are still valid, but with diminished performance. In this paper, we present a new clustering approach based on multilevel paradigm called multilevel clustering, that allows to divert the complexity of calculation and execution period of data mining on very large scale. The developed algorithm have been implemented on three public benchmarks to test the effectiveness of the multilevel clustering approach. The numerical results have been compared to those of the simple k-means algorithm. As foreseeable, the multilevel clustering outperforms clearly the basic k-means on both the execution time and success rate that reaches 100 % while increasing the number of data. © Springer International Publishing Switzerland 2013.