New scalable varied density clustering algorithm for large datasets

Finding clusters in data is a challenging problem especially when the clusters are being ofwidely varied shapes, sizes, and densities. Herein a new scalable clustering technique whichaddresses all these issues is proposed. In data mining, the purpose of data clustering is toidentify useful patterns in the underlying dataset. Within the last several years, manyclustering algorithms have been proposed in this area of research. Among all these proposedmethods, density clustering methods are the most important due to their high ability to detectarbitrary shaped clusters. Moreover these methods often show good noise-handlingcapabilities, where clusters are defined as regions of typical densities separated by low or nodensity regions. In this chapter, we aim at enhancing the well-known algorithm DBSCAN, tomake it scalable and able to discover clusters from uneven datasets in which clusters areregions of homogenous densities. We achieved the scalability of the proposed algorithm byusing the k-means algorithm to get initial partition of the dataset, applying the enhancedDBSCAN on each partition, and then using a merging process to get the actual natural numberof clusters in the underlying dataset. This means the proposed algorithm consists of threestages. Experimental results using synthetic datasets show that the proposed clusteringalgorithm is faster and more scalable than the enhanced DBSCAN counterpart.