论文部分内容阅读
核函数、惩罚因子、核参数是影响支持向量数据描述(SVDD)分类方法分类效果的重要因素。研究了多核支持向量数据描述(MKSVDD)分类方法,给出了多核支持向量数据描述分类方法的实现步骤,基于banana数据集分析了惩罚因子和核参数对分类效果的影响,重点讨论了多核函数的权值对支持向量数据描述边界分布的影响。仿真实验结果表明,与单核支持向量数据描述分类方法相比较,多核支持向量数据描述分类方法的分类效果更佳,为实际应用时参数的选择提供了参考。
Kernel functions, penalty factors, and kernel parameters are important factors that influence the classification of SVDD classification methods. The classification method of multi-kernel support vector data description (MKSVDD) is studied. The implementation steps of multi-core support vector data description classification method are given. Based on the banana data set, the influence of penalty factor and kernel parameter on the classification result is analyzed. The Influence of Weight on Describing Boundary Distribution by Support Vector Data. The simulation results show that compared with single-kernel support vector data classification method, multi-core SVR data classification method can be better classified, which provides reference for parameter selection in practice.