A Competitive Clustering on Bigdata

The vast increase in the volume of data has created a need for new applications and algorithms to quickly analyse the large scale of data. Many of the cluster analysis techniques like K-Means are used to compute the data in distributed systems, its accuracy depends on the initial seeding of centroids. The improvisation of K-Means algorithm shows good initial seeding but it suffers with the serial nature i. e. , it takes long time on large data sets. In this paper we propose a new algorithm with Map Reduce implementation to address the above problems. Our algorithm provides parallel processing of data by dividing it into number of subsets using Hadoop with MapReducing methods. Our work provides good initial centers in a less time and also produces fast and accurate cluster analysis on large scale data.