一种RBF核SVM的参数选择方法
Method for Selecting Parameter of SVM with RBF Kernel
讨论了参数误差惩罚参数、核参数对RBF核SVM推广能力的影响，分析了误差惩罚参数、核参数取值的合理范围。在此基础上，对双线性网格搜索法进行了改进，实验证明，改进的方法能够排除误差惩罚参数、核参数取值过大、不稳定的情况，并且预测精度能够达到双线性网格搜索法的高度，具有较好的性能。
his paper studied the influences of the error penalty parameter and the kernel parameter on support vector machine\\\\\\\\\\\\\\\'s generalization ability and analyzed the reasonable range of the error penalty parameter and the kernel parameter. Bilinear-grid method was improved to confirm this two parameters. Experiment results show that the improved method has similar predict accuracy and better performance compared to bilinear-grid search method.
高瞻、闫志刚、丁允静
计算技术、计算机技术
支持向量机RBF核推广能力双线性网格搜索法
support vector machineRBF kernelgeneralization abilitybilinear-grid search method
高瞻,闫志刚,丁允静.一种RBF核SVM的参数选择方法[EB/OL].(2010-05-26)[2025-08-16].http://www.paper.edu.cn/releasepaper/content/201005-659.点此复制
评论