Particle Swarm Optimization Algorithm Based k-means and Fuzzy c-means clustering

Data mining is the process of extracting hidden patterns from huge data. Among the various clustering algorithms, k-means is the one of most widely used clustering technique in data mining. The performance of k-means clustering depends on the initial clusters and might converge to local optimum. K-means does not guarantee the unique clustering because it generates different results with randomly chosen initial clusters for different runs of kmeans. In addition, the performance of fuzzy c-means depends on membership matrix [μ] value and might not guarantee the unique clustering. This paper explains the application of evolutionary algorithm namely Particle Swarm Optimization and Entropy based fuzzy clustering for identifying the initial centroids for enhancing the performance of both k-means and fuzzy c-means clustering. Keywords—k-means clustering, fuzzy c-means clustering, cluster initialization, Particle swarm optimization, Entropy based fuzzy clustering