摘要
考虑装备保障能力评估实际中获得的数据被污染的问题,在最小二乘支持向量机算法的基础上对其进行改进,提出了一种鲁棒学习的最小二乘支持向量机方法,使得所建模型可以避免奇异点影响。通过仿真验证了该方法拟合精度高、鲁棒性强,将其应用到装备保障能力评估中提高了评估精度,具有较好的推广性。
Considering the contamination of data obtained in equipment support capability evaluation,the Least Squares Support Vector Machine（LS-SVM） was improved,and a method of robust LS-SVM was proposed to make the established model be free of the influence of singular points.Simulation was made and the result showed that that the improved method has fine fitting accuracy and high robustness,and can improve the evaluation precision when it is used in the equipment support capability evaluation.
出处
《电光与控制》
北大核心
2011年第11期53-56,共4页
Electronics Optics & Control
关键词
装备保障
评估模型
最小二乘法
支持向量机
鲁棒性
equipment support
evaluation model
least square method
support vector machine
robustness
