A local network neighbourhood artificial immune system for data clustering

The artificial immune system (AIS) is inspired by the functioning of the natural immune system. There are different theories with regards to the organisational behaviour of the natural immune system. One of these theories is the network theory. In this paper a novel network based AIS model is proposed. The proposed Local Network Neighbourhood Artificial Immune System (LNNAIS) is inspired by the network topology of lymphocytes to learn the antigen structure from one another. LNNAIS has a different interpretation of the network theory compared to existing network based AIS models. LNNAIS uses a concept of an artificial lymphocyte (ALC) neighbourhood to determine the network links between the ALCs. The purpose of this paper is to provide a proof of concept that an artificial lymphocyte (ALC) neighbourhood can cluster data in a dynamic environment. LNNAIS only requires one pass through the training data of antigen patterns for clustering.