Dynamic and Optimized Prototype Clustering for Relational Data based on Multiple Prototypes

Abstract

In artificial intelligence related applications such as bio-medical, bio-informatics, data clustering is an important and complex task with different situations. Prototype based clustering is the reasonable and simplicity to describe and evaluate data which can be treated as non-vertical representation of relational data. Because of Barycentric space present in prototype clustering, maintain and update the structure of the cluster with different data points is still challenging task for different data points in bio-medical relational data. So that in this paper we propose and introduce A Novel Optimized Evidential C-Medoids (NOEC) which is relates to family o prototype based clustering approach for update and proximity of medical relational data. We use Ant Colony Optimization approach to enable the services of similarity with different features for relational update cluster medical data. Perform our approach on different bio-medical related synthetic data sets. Experimental results of proposed approach give better and efficient results with comparison of different parameters in terms of accuracy and time with processing of medical relational data sets. Index Terms: Data clustering, multiple prototypes, artificial intelligence, and prototype based clustering, c-medoids and Ant Colony Optimization (ACO).