A Proposed Modification Over K-Means Algorithm for Data Clustering and Pattern Classification

Abstract

Artificial Intelligence are extensively used nowadays to carry out tasks like prediction, pattern matching, detection, which are carried out by the persons who acquire a high level of expertise and knowledge in medical field. Supervised Networks  
 are widely applied for the diagnosis and prediction of the Diabetes and other diseases. The Artificial Intelligence is applied to conduct tasks require machine interference and automation like interpretation and analysis of image etc. They are used along with classification techniques to forecast patterns, behaviors and trends for modeling intense medical problems like Cancer, Blood Pressure, Heart Disease, and in our research work for Diabetes Type-2.

The work that discuss in this dissertation report focuses on the effectual diagnosis  and diagnosis of diseases like Diabetes by using data clustering and classification techniques, thereby modifying these algorithms to enhance their performance. In this work Learning Vector Quantization and K-Means Clustering algorithm simulation has been done, and function replacement is done.  The results showed that it enhances the performance of the output generated by these algorithms.