摘要
针对现代战场信息化程度的不断提高,电磁环境日趋复杂,侦查目标难以准确地识别情况,提出了运用支持向量机多分类器对军事侦查目标进行有效识别。结构风险最小化地支持向量机分类方法是小样本情况下统计机器学习的经典,具有速度快、泛化能力强等特点。用该算法建模军事目标的识别问题,达到了较高的识别准确率。所以应用在对侦查目标的识别上具有良好的效果,在军事应用上有较广阔的前景。
The informatization of modern battlefield is unceasing increasing its electromagnetic environment is complex day by day,as a result,the target recognition is difficult to be accurate in target reconnaissance. This paper proposed a method that can effectively recognize the target through the use of multi-classification support vector machine. The structure risk of this method is smallest,have fast spped and strong generalization. This method is used to model the recognition of military, and obtained higher accuracy of recognition
出处
《火力与指挥控制》
CSCD
北大核心
2009年第8期97-100,共4页
Fire Control & Command Control
关键词
支持向量机
二叉树多分类器
军事目标识别建模
support vector machine ,multi-sorters of binary tree ,military target recognition modeling
作者简介
唐克（1962-），男，安徽东至人，副教授，硕士生导师，主要研究方向：武器系统分析与仿真。
