Improved density based algorithm for data stream clustering

In recent years,clustering methods have attracted more attention in analysing and monitoring data streams. density-based techniques are the remarkable category of clustering techniques that are able to detect the clusters with arbitrary shapes and noises. however,finding the clusters with local density varieties is a difficult task. for handling this problem,in this paper,a new density-based clustering algorithm for data streams is proposed. this algorithm can improve the offline phase of density-based algorithm based on minpts parameter. the experimental results show that the proposed technique can improve the clustering quality in data streams with different densities. © 2015 penerbit utm press. all rights reserved.