Dynamic clustering using support vector learning with particle swarm optimization

This paper presents a new approach to the support vector learning for dynamic clustering based on particle swarm optimization. Support vector clustering requires solving a constrained quadratic optimization problem. This problem often involves a matrix with an extremely large number of entries, which make off-the-shelf optimization packages unsuitable. Several methods have been used to decompose the problem, of which many require numeric packages for solving the smaller subproblems. This paper gives an overview of the support vector clustering algorithm. Particle swarm optimization is discussed as an alternative method for solving a support vector clustering's quadratic programming problem. Experimental results illustrate the convergence properties of the algorithms.