Analysis and Optimization of Data Classification using K-Means Clustering and Affinity Propagation Technique

Abstract Twitter provides an enormous platform to perform analysis on data consisting of events, trends, personalities. It enables to determine the inclination, likes of the people in real time independent of size. There several techniques to retrieve the data and the most efficient technique to retrieve the data is the clustering technique. There are many approaches in clustering to group and analyze the data. This paper provides an overview on various algorithms and their effectiveness in determining the trending pulses efficiently. Once the data are clustered, they could be classified based on the topics for real time analysis on the huge collection of data set which is very dynamic. In this paper classification of data is performed and analyzed to determine the flaws. The classification is again performed on the same dataset using an optimized technique and analysis is performed on the clustering of the data. Keywords: K-means Clustering, Affinity Propagation Technique, Centroid, Euclidian Distance