基于图论的支持向量机核函数选择

支持向量机是以核方法为核心的一类机器学习方法,不同的核函数具有各异的度量特征,故核函数的选取对支持向量机性能有着重要的影响并成为支持向量机研究中需要解决的核心问题之一。目前,核函数的选择大多数存在一定的随意性和局限性。为此,利用图论构建样本数据无向图的邻接矩阵,估计其本征维数,有指导性地选择支持向量机核函数类型,提高支持向量机的泛化能力。实例仿真和相似性对比验证了该方法的有效性。

Selection of SVM kernel function based on graph theory

Support vector machine is a kind of machine learning method with kernel method as the core,different kernel functions have different measurement characteristics,so the selection of kernel function has an important influence on the performance of SVM and becomes one of the core problems to be solved in SVM research.At present,most of the choice of kernel function has certain randomness and limitation.Therefore,the adjacency matrix of the undirected graph of sample data was constructed using graph theory and its intrinsic dimension was estimated,the kernel function of the support vector machine was guided to improve the generalization ability of the support vector machine.The simulation of the example and the comparison of the similarity show that the proposed method is effective and feasible.