Immune Based Clustering for Medical Diagnostic Systems

It has recently been shown that Artificial Immune Systems (AIS) can be successfully implemented as biologically inspired systems. The Artificial Immune Network (AIN) is one of the computationally intelligent systems that are inspired by the processes of the immune system. Many algorithms are used to exploit the immune system's characteristics of learning and memory. The aiNet is one of such AIS algorithms with excellent performance on elementary clustering tasks. This paper proposes the use of two clustering techniques (DBSCAN and K-MEANS) in combination with aiNet algorithm for medical diagnosis. Two standard data sets are used to evaluate the clustering performance. The results indicate that both clustering techniques produced closely related outcomes under the usage of aiNet algorithm. However, K-means showed higher accuracy and better percentage of total clustering than DBSCAN for both data sets.