基于影响因素主成分分析的航材消耗预测模型  

Prediction Model of the Aviation Materials Consumption Based on PCA of Influencing Factors

作　　者：张梁[1] 崔崇立[2] 贲徐伟 辛昱[1] 曹亮[4] 

机构地区：[1]空军勤务学院学员一大队二队 [2]空军勤务学院航材管理系 [3]中国人民解放军94906部队 [4]中国人民解放军94783部队

出　　处：《兵工自动化》2016年第8期50-54,共5页Ordnance Industry Automation

摘　　要：为了确定航材消耗预测各影响因素的影响程度,提高航材消耗规律分析及其预测的准确性,提出一种基于影响因素主成分分析与广义线性回归的航材消耗预测模型。对航材消耗影响因素进行定性、定量分析,通过主成分分析筛选变量,对影响因素数据进行降维,得到最佳的影响因子集合;然后通过广义线性回归构建航材消耗方程,并验证其拟合优度,最后进行消耗预测。分析结果表明:该模型降低了影响因子相关性,减小了回归分析难度,提升了回归分析的精度,提升了航材消耗预测模型的准确度。To define the multi-factor affecting degree of consuming prediction of aviation materials and improve the accuracy of consumption analysis and prediction, we proposed a method for predicting the consumption of aviation materials based on principal component analysis（PCA） of influencing factors and generalized liner regression. Firstly, quantitatively and qualitatively analysis the influence factors of aviation materials consumption. Then, screen out the main factors by PCA and make it dimensionality reduction to get the best collection of variables. Finally, use the generalized linear regression to build the air material consumption equation, and verify its goodness-of-fit, and finally make consumption forecast. The results show that the model reduces the impact factor correlation, and the difficulty of regression analysis, and improves the precision of regression analysis, and the accuracy of the prediction model of aviation material consumption.

关 键 词：航材消耗 主成分分析 广义线性回归 

分 类 号：TP391.92[自动化与计算机技术—计算机应用技术]