Research And Application Of Two Kinds Of Coupling Algorithm For Support Vector Machine Parameter Optimization

Support vector machine to fuse the optimal hyper plane,soft space and nuclear idea and so on,SVM is one of the most effective methods to solve the problem of high dimensional nonlinear classification.The kernel parameter and penalty parameter have great influence on the classification accuracy and generalization ability of SVM.The algorithm is a kind of algorithm,which combines the two features,and combines the one features to optimize the parameters of SVM.The ant colony algorithm is easy to premature convergence and stagnation phenomenon,fall into the local extremum problem,proposed an improved ant colony optimization algorithm(IACO).The algorithm s pheromone update strategy defines a new dynamic adaptive global pheromone updating function;aiming at the basic artificial bee colony algorithm is easy to fall into a long-term stagnation in global search ability strong development ability is insufficient,and chaotic search algorithm has very good and ergodicity,the chaotic detection mechanisms(artificial bee colony algorithm,the algorithm with chaotic detection mechanisms get new nectar to iteration,the algorithm can jump out of local optimum.The improved two kinds of algorithms are applied to support vector machine parameter optimization.The IACO-SVM model and IABC-SVM model are used to carry out numerical experiments.The results show that the IACO-SVM model has strong average classification accuracy and better algorithm stability.IABC-SVM model can improve the accuracy of support vector machine classification and reduce the invalid iteration.The algorithm can effectively jump out of local extremum.