A Clustering Method Using Simplified Swarm Intelligence Algorithm

Abstract

The main aim of clustering is to represent large datasets by a fewer number of segments or partitions. It brings simplicity in modeling data, knowledge discovery and data mining. A family of bio-inspired algorithms, well-known as Swarm Intelligence (SI) has recently emerged that can successfully be applied to a number of real world clustering problems. This paper presents a Bird flocking algorithm that uses the concepts of a flock of agents (birds) moving together in a complex manner with simple local rules. Our algorithm dynamically creates and visualizes groups of data. Each agent/bird representing one data, move with the aim of creating homogeneous groups of data in a 2D environment. This simulation is carried by emulating the natural flocking behavior of birds requiring behavioral rules for cohesion, separation, alignment and avoidance of birds belonging to different community. The birds effectively “think” for themselves and move in a self-organizing manner allowing study of their behavior.

Keywords: Clustering, Swarm Intelligence, Flocking, Collision, Separation, Alignment, Avoidance.