HYBRIDIZATION OF CLUSTERING ALGORITHM FOR BETTER OPTIMIZATION

In data tunneling and data analyses, clustering plays a vital role. Clustering is a technique wherein the resultant groups formed consists of data which are like one another. Cluster analysis groups testimony based on correlation and diversity among the data elements. These groups formed are called as clusters and it is an unsupervised approach. There are distinct algorithms which consider the attributes of data and the crunch numbers to form clusters from the info. Based on the behavior of the algorithm the centroids are preferred naturally by algorithm or the user can define it. The preferred algorithm for clustering is K-Means which splits up the data based on the mark of compactness, but it also has some demerits like falling in local optimum. So, in order to avoid, that another algorithm which can be used is fuzzy clustering algorithm (FCM). To get hold of fuzzy patterns as a turnout method called fuzzy clustering is utilized. FCM also has another face which describe that the Euclidean distance measures can unevenly weight underlying factors. Getting uplifted from the decorum of birds, particle swarm optimization (PSO) is a worldwide enhancement process. PSO is popularly used in many cluster analysis issues. So, in order to make algorithm gives better results, we are bringing together two algorithms which will take advantage of twain design. This works set forth the hybrid combination of K-means and PSO called as Hybrid-PSO. The motive responsible for linking 2 algorithms is that it gives preferable results in terms of speed and proves to be strong clustering algorithms which will be shell out fitter optimization.