Heuristic Algorithm in the Application of SVM Parameter Optimizing

The selection of SVM classifier parameters is the key factor that influence the efficiency of the classifier, CV optimization method need to traverse all the parameters points within the grid, to find the global optimal solution, high time complexity of the algorithm. Heuristic algorithm can need not traverse all parameters points in the grid, and find the global optimal solution. The article describes the two kinds of optimization parameters of the heuristic algorithm, they are genetic algorithm and particle swarm algorithm, heuristic algorithm to reduce the time complexity of classifier, to improve the classification efficiency.