仿真实验与SVM相结合的数据分类方法研究  

Study on Data Classification Method Based on Simulation Experiment and SVM

作　　者：祝曙光[1,2] 胡晓峰[1] 司光亚[1] 杨镜宇[1] 张明智[1] 罗批[1] 郭永辉[1] 

机构地区：[1]国防大学信息作战与指挥训练教研部,北京100091 [2]总装备部工程设计研究总院,北京100028

出　　处：《系统仿真学报》2010年第3期761-764,共4页Journal of System Simulation

基　　金：国家863高技术计划(2006AA01Z337);博士后科学基金(2005038053)

摘　　要：首先提出了仿真实验与SVM相结合的数据分类方法;其次,利用该方法成功地进行了若干数据分类实验,证明了该方法能够高效、优质地完成体系规划与优化过程中的数据分类任务。此外,还探讨了二分类支持向量机的推导过程,研究了利用仿真实验为二分类支持向量机准备实验样本的具体方法。Data classification method based on simulation experiment and support vector machine （SVM） was put forward. And several data classification experiments have been carried out successfully with this method, which indicates that this new method can be used to fulfill the task of data classification in System-of-Systems （SOS） programming and optimization process effectively. In addition, the derivation process of SVM for data set with two different data types was given and the SVM experiment sample preparation method based on simulation experiments was also discussed.

关 键 词：体系规划与优化 数据分类 支持向量机 仿真实验 

分 类 号：E911[军事] TP391.9[自动化与计算机技术—计算机应用技术]