基于灰色特征加权支持向量机的火炮质量评估  

Quality evaluation of artillery based on grey feature weighted support vector machine

作　　者：胡胜海 何蕾 徐鹏[1] 杨奇 金锦花 

机构地区：[1]哈尔滨工程大学信息与通信工程学院,黑龙江哈尔滨150001

出　　处：《应用科技》2011年第1期21-25,共5页Applied Science and Technology

基　　金：国防科学技术工业委员会基础研究基金资助项目(40108060302)

摘　　要：针对火炮质量评估中现场数据少,传统评估方法过多地依赖人的主观因素的问题,提出了一种基于灰色特征加权支持向量机(GFWSVM)的火炮质量评估方法.针对评估因素重要度的不同,引入特征加权的概念.采用灰色关联分析求取因素特征权重,避免了评估结果被一些弱相关因素所支配.通过对火炮质量评估实例仿真计算,表明该方法评价结果可靠,可操作性强,说明了该方法的有效性.An approach to evaluate artillery quality based on grey feature weighted support vector machine （GFSVM） is presented to solve the problem of lacking field data and weaken the subjective factors which are heavily depended on in traditional evaluation method in artillery quality evaluation. The method of grey correlation analysis is used to obtain the feature weight so as to avoid the evaluate result to be dominated by the weak correlation factors according to the different importance of each evaluating factor. Finally, a case of artillery quality evaluation is used to testify the method. The result shows that the method proposed is credible and effective, and can be applied easily.

关 键 词：支持向量机 特征加权 灰色关联分析 质量评估 火炮 

分 类 号：TH12[机械工程—机械设计及理论]