Kernel based automatic clustering using modified particle swarm optimization algorithm

This paper introduces a method for clustering complex and linearly non-separable datasets, without any prior knowledge of the number of naturally occurring clusters. The proposed method is based on an improved variant of the Particle Swarm Optimization (PSO) algorithm. In addition, it employs a kernelinduced similarity measure instead of the conventional sum-ofsquares distance. Use of the kernel function makes it possible to cluster data that is linearly non-separable in the original input space into homogeneous groups in a transformed highdimensional feature space. Computer simulations have been undertaken with a test bench of five synthetic and three real life datasets, in order to compare the performance of the proposed method with a few state-of-the-art clustering algorithms. The results reflect the superiority of the proposed algorithm in terms of accuracy, convergence speed and robustness. Categories and Subject Descriptors I.2.2 [Automatic Programming], I.2.6 [Learning], I.5.3 ...

• Type: Conference
• Year: 2007
• Where: GECCO
• Authors: Ajith Abraham, Swagatam Das, Amit Konar
• Keywords: Conventional Sum-ofsquares Distance, GECCO 2007, Particle Swarm Optimization, State-of-the-art Clustering Algorithms