Efficient Graph Clustering Algorithm and its use in Prediction of Students Performance

Abstract

Big data refers to its massive unstructured information that comes from the increasing use of technologies. Analyzing and extracting this information is a great challenge including data clustering. In order to overcome this challenge and to maintain its scalability and performance, an efficient clustering algorithm and a framework is necessary. So, this paper presents a refined clustering algorithm based on power method. The algorithm is made to run on a hadoop cluster. We also present case studies to demonstrate the difference among the results of clusters. A detailed study is made on the student's academic performance of a particular department in our college focusing on the factors that affect their placement.