二叉树多分类SVM在目标分群中的应用

Application of Multiclass SVM Based on Binary Tree in Target Grouping

为了解决目标分群问题,在研究目标关键属性的基础上,提出一种基于二叉树多分类支持向量机(SVM)的目标分群方法。介绍了基于统计学习理论的支持向量机方法的基本原理和算法本身的理论优势,由于支持向量机的本质是解决二分类问题的,因此如何建立支持多分类的支持向量机是研究的关键。采用基于二叉树的多分类支持向量机算法,建立了解决目标分群问题的算法模型,将分类器分布在各个节点上,从而构成了多分类的支持向量机。将算法进行软件实现,并利用模拟数据对支持向量机进行了训练,调整支持向量机参数,得到了较好的分群结果,证明了该方法的有效性。

Support Vector Machines （ SVMs） is originally designed for linear binary pattern recognition. Many multiclass SVMs are introduced briefly in the paper.Then the method of multiclass SVM based on binary tree is designed after analyzing the key issues of target grouping.The model of the multiclass SVM algorithm is designed.And the algorithm is realized by coding.The algorithm parameters and weighting factors are adjusted by simulation experiment. The result of simulation validates the feasibility and correctness of the grouping algorithm.