Quality Based Analysis of Clustering Algorithms using Diabetes Data for the Prediction of Disease

Abstract

Clustering is the popular fundamental investigative performance analysis technique commonly used in various applications. The majority of the clustering techniques proved their effectiveness in finding lot of solutions for a variety of datasets. With the aim of test its performance and its clustering qualities are easy to implement by partition based clustering algorithms. The clustering algorithms k-Means and k-Medoids are used to analyze the diabetic datasets and to predict the diseases in this research work. Around 15000 diabetic patient’s consequential final bio-chemistry prescription are taken for the diabetes identification. With number of times executed the run time of the algorithms are compared from the different clusters. Based on their performance the first-rate algorithm in each class was found out.. The best suitable algorithm is suggested for the prediction of diabetes data in this work.