摘要
针对军用无人机研制费用的影响因素多且复杂,样本数据又相当匮乏的问题,充分利用偏最小二乘回归方法在处理小样本数据,变量多重相关性等方面的独特优势,给出了基于偏最小二乘回归方法的预测模型和军用无人机研制费用的算法步骤.用文献[2]中的数据进行了检验,并与逐步多元回归、反向传播(BP,Back Propagation)和径向基函数(RBF,Radial Basis Function)神经网络法进行了比较,结果表明了偏最小二乘模型的实效性.
The basic model was formulated based on the partial least-squares regression method,which has some prominent advantages in dealing with small sample and variables with multiple correlations,and the arithmetic steps of the development cost prediction of the military unmanned air vehicle were given,for there are some complex factors such as maximum take-off mass,cruise velocity,flight altitude and so on affecting the development cost of the military unmanned air vehicle and few sample data of the military unmanned air vehicle.The data in reference ／ was tested and compared with step multivariate regression,back propagation and radial basis function neural network method.The results show that the partial least-squares regression model is practical and efficient.
出处
《北京航空航天大学学报》
EI
CAS
CSCD
北大核心
2010年第6期667-670,共4页
Journal of Beijing University of Aeronautics and Astronautics
关键词
军用无人机
研制费用
偏最小二乘回归
神经网络
military UAV（unmanned air vehicle）
development cost
partial least-squares regression
neural network
