论文部分内容阅读
目的针对支持向量机(support vector machine,SVM)的参数选择的重要性,研究一种新的参数优化方法。方法介于蝙蝠算法的模型简单、全局搜索能力强等特点。本文提出基于蝙蝠算法(BA)的SVM参数优化方法,对SVM的惩罚参数和核参数进行优化。结果通过8个UCI标准数据库集的Matlab仿真实验,验证了算法的有效性和可靠性。结论本文方法搜索的最优参数较大地提高了SVM的分类精度,加强了SVM的学习和泛化能力,是一种有效及稳定的支持向量机参数优化方法。
Aiming at the importance of parameter selection of support vector machine (SVM), a new parameter optimization method is studied. The method lies in that the model of the bat algorithm is simple and has strong global search capability. In this paper, the method of SVM parameter optimization based on bat algorithm (BA) is proposed, and the penalty parameters and kernel parameters of SVM are optimized. Results The Matlab simulation of eight UCI standard database sets verified the effectiveness and reliability of the proposed algorithm. Conclusion The optimal parameters of this method search greatly improve the classification accuracy of SVM and strengthen the learning and generalization ability of SVM. It is an effective and stable parameter optimization method for SVM.