Search

Support vector machine (SVM) parameter optimization selection is an important research direction, but there is still no systematic theory to guide the selection of the SVM parameters. Since optimizing the SVM parameters by the artificial fish-swarm algorithm tends to fall into the small neighborhood of the approximate optimal solution, we design the AFMC algorithm for the SVM parameter optimization. At the early stage, we use the better parallel optimization performance of the fish-swarm algorithm to quickly gain the approximate optimal solution. Then we use the MonteCarlo algorithm for local searching to achieve a quick and effective strong-approximate optimal solution. The numerical experiments show that the proposed algorithm has better classification performance and faster searching speed, and it is effective and feasible in the SVM parameter optimization. [ABSTRACT FROM AUTHOR]

Solving a large-scale convex quadratic programming problem is the core of the support vector machine. An equivalence complementarity problem can be obtained based on an amended model of the support vector machine(SVM), therefore we propose a non-monotonic trust region algorithm for solving the complementarity problem based on the Fischer-Burmeister complementarity function. The new algorithm need not compute any Hesse or the inverse matrix, thus reducing the amount of computational work. Global convergence of the algorithm is proved without any assumptions. Numerical experiments show that the running speed of the algorithm is faster than that of the LSVM algorithm and the descent algorithm when solving large-scale nonlinear classification problems and thus it provides a feasible method for solving SVM. [ABSTRACT FROM AUTHOR]