基于KPCA和MSVM的雷达装备作战能力评估方法研究

Research on the Evaluation Method of Radar Equipment Combat Capability Based on KPCA and MSVM

针对传统降维方法以及综合评估方法不能满足日益复杂的装备作战能力评估需求这一突出矛盾，利用KPCA在非线性降维方面、MSVM在非线性综合评估方面的优势，提出了KPCA与MSVM复合的新型装备作战能力评估方法。基于该复合算法，利用MATLAB分别构造了KPCA降维模型与MSVM综合评估模型。以雷达装备这一典型电子信息系统为评估对象，建立完整的作战能力评估指标体系。选取数据，运用构建的评估模型进行综合评估。比较运用KPCA与MSVM复合、PCA与MSVM复合及MSVM独立评估三种不同评估方法在进行雷达装备作战能力评估时的结果，表明运用KPCA与MSVM复合的算法，在降维提高运算效率的基础上，分类准确率相对于直接运用MSVM进行评估时能够达到80％，而同条件下PCA与MSVM复合算法的分类准确率仅能达到30％。因此，运用KPCA和MSVM复合的方法进行评估时，不仅能够实现降维提高运算效率，而且评估结果准确性较高，适于进行装备作战能力评估。

Aimed at the conflict that traditional dimensionality reduction method can ＇ t meet the demand of complex equipment combat capability evaluation, and based on the advantage of KPCA and MSVM in nonlinear dimensionality reduction, a new equipment combat capability eval- uation method is advanced, which combined the KPCA and MSVM together. Based on this combined method, and using Matlab method, the KPCA dimension reduction model and MSVM general evaluation model are constructed respectively. Taking radar equipment electronic information system as an example, the corresponding evaluation index system is built. The constructed evaluation model is adopted to do the general evaluation with data. The results of such three kinds of combination which including KPCA-MSVM, PCA-MSVM and uniquely MSVM are compared. And the result shows that based on the improvement of computing efficiency, the classify accuracy of the combined method KPCA-MSVM can reach 80%, compared with the other two methods. Under the same condition, the classify accuracy of PCA-MSVM only has 30%. So we concludes that the combined method KPCA-MSVM can not only accelerate the computing speed, but also has relatively high classify accuracy, and is a suitable dimensionality reduction method for the equipment combat capability evaluation.