Clustering is a helpful cheme that organizes automatically a large amount of unstructured data into clusters that comprises of structured data so that it has better similarity as compared to the unstructured document in other clusters. The use of document clustering increasing day by day due to its applicability in different area like web mining, search engine and to retrieve information. It this paper, we present K-mean clustering technique which is used to organize the unstructured data into structured data known as clusters. Then Cosine similarity index is used in order to determine the similarity measures among clusters. At last the metrics like FAR, accuracy, FRR and entropy is evaluated in MATLAB simulation tool. We made use of K-mean clustering to affectively process the segmentations and categorizations. By using the K-mean clustering, we are going to proceed with behavioral segmentation on entropies and categorize the clustering records to be evaluated with simulation tools. Making the centroid update in this study is very accurate and supported the simulations very well.
Author(s): Swathi Muppalaneni