A Scheme of Function Approximation Based on SVM & NSGA II

A scheme of function approximation based on support vector machine (SVM) and nondominated sorting genetic algorithm II (NSGAII) is proposed. By combining structural risk minimization principle of SVM, computational complexity of kernel function and function approximation error norm as a multi-objective optimization problem, parameters and kernel function for function approximation are automatically chosen by adopting NSGAII. Simulation shows the scheme has higher precision and less complex kernel function than conventional SVM function approximation