摘要
为预测军用飞机维修费用,选取军用飞机维修费用预测指标,使用收集的训练样本,建立军用飞机维修费用预测的随机森林模型,并预测军用飞机维修费用;将该模型预测结果与支持向量机回归(SVR)模型和BP神经网络模型的预测结果进行对比。研究表明:采用RF模型所预测的实例的军用飞机维修费用误差为6.9%,低于SVR模型的误差7.6%,BP神经网络的误差8.4%;可见RF模型具有较高的预测精度,能够相对有效地对合同经费进行预测。
In order to predict the maintenance cost of military aircraft,a stochastic forest model is established to predict the maintenance cost of military aircraft by selecting the prediction index of military aircraft maintenance cost and using the collected training samples;the prediction results of this model are compared with those of support vector machine regression(SVR)model and BP neural network model.Research shows:the error of military aircraft maintenance cost predicted by RF model is 6.9%,which is lower than 7.6%of SVR model and 8.4%of BP neural network;it can be seen that RF model has high prediction accuracy and can be used to predict contract funds relatively effectively.
作者
王杰
王勤
崔凯
韩戈白
Wang Jie;Wang Qin;Cui Kai;Han Ge-bai(China Electronics Technology Group Corporation 28 Research Institute,Jiangsu Nanjing 210000;Air Force Equipment Department,Beijing 100010)
出处
《电子质量》
2020年第12期101-107,共7页
Electronics Quality
关键词
随机森林模型
军用飞机维修费用
数值分析
预测精度
random forest model
military plane maintenance costs
numerical analysis
predictive accuracy