基于马氏距离的舰船装备修理价格组合预测

Maintenance Fees Forecast for Warship Equipment Based on Mahalanobis Distance

为更加准确地确定舰船装备全寿命费用,提高装备经费的使用效益,针对我国舰船装备修理价格影响因素多、不确定性大的特点,提出采用组合预测方法对其进行预测建模;针对基于欧氏距离的组合预测方法不能处理用于组合的单项预测模型之间的相关性问题,提出采用马氏距离作为建模基础,建立基于马氏距离的舰船装备修理价格组合预测模型。应用结果显示,基于马氏距离的组合预测要比基于欧氏距离的组合预测以及各单项预测具有更好的稳定性,预测风险更小。

The determination of the maintenance fees for warship equipment involves many factors and huge uncertainty. To confirm the life cycle cost of warship equipment more exactly, the forecast combination method is applied in this paper to establish the forecast model for warships, which ensures efficient use of equipment funding. As there is a linear correlation among different individual models, the forecast combination model based on Euclidean distance may be poor in performance. Therefore, Mahalanobis distance is proposed as a substitution. Results show that the proposed method, compared with the forecast combination based on Euclidean distance and individual forecast models, is less risky and more reliable.