基于递阶偏最小二乘回归的飞机采购价格预测  

The Purchasing Price Prediction of Aircraft based on Hierarchical Partial Least Squares Regression

作　　者：王永杰[1] 王礼沅[1] 张恒喜[1] 郭基联[1] 

机构地区：[1]空军工程大学工程学院,西安710038

出　　处：《火力与指挥控制》2010年第10期98-101,共4页Fire Control & Command Control

摘　　要：分析了飞机采购价格预测建模样本数据少、价格驱动因子众多的特点,考虑到递阶偏最小二乘回归(Hierarchical Partial Least-Squares Regression,Hi-PLS)方法在变量规模巨大情形下进行回归建模的优势,应用Hi-PLS对飞机采购价格进行预测。以战斗机采购价格预测为例进行了研究,首先对战斗机采购价格驱动因子进行分组,然后应用Hi-PLS对分组后的价格驱动因子进行回归,建立采购价格预测模型。实例表明,在飞机采购价格预测方面,采用递阶偏最小二乘回归预测更能体现价格与飞机性能参数之间的关系。Analyzing small samples and various price drive factors of the aircraft＇s purchasing price and considering the advantages of Hierarchical Partial Least-Squares Regression（Hi-PLS） in regressing in the condition of large number＇s variables,an application of Hi-PLS to the aircraft＇s purchasing price prediction is proposed.Using the fighter plane＇s purchasing price prediction as research example,after the price drive factors of the fighter plane＇s purchasing price are grouped,Hi-PLS is applied to the regress of the grouped price drive factors,and the fighter plane＇s purchasing price prediction model is established.The calculated results in the example show that,Hierarchical Partial Least-Squares Regression model can reflect better the correlation between purchasing price and performance parameters of military aircrafts.

关 键 词：飞机 采购价格 递阶偏最小二乘回归 价格预测模型 价格驱动因子 

分 类 号：V37[航空宇航科学与技术—航空宇航推进理论与工程] O241.5[理学—计算数学]