An improvement of DENCLUE algorithm for the data clustering
Abdellah Idrissi, Hajar Rehioui, Abdelquoddouss Laghrissi, Sara Retal · 2015
Classification is one of important tasks in the Data Mining field. It aims to merge the similar data into a group. In this context, several methods of classification have been proposed in literature. DENCLUE (DENsity-based CLUstEring) is one of the most effective unsupervised classification methods, that allows to classify voluminous data. This method is based on the concept of density and the Hill Climbing algorithm. The Hill Climbing helps in the crucial phase of the reconstruction of the classes. In this paper, our ultimate goal is to increase the performance of DENCLUE in terms of better classification and execution time. For this purpose, we propose to replace the Hill Climbing firstly by the Simulated Annealing (SA) and secondly by a Genetic Algorithm (GA). We tested these two approaches on datasets extracted from the literature. The experimental results showed the performance of our proposals.