摘要
针对军用无人机研制费用预测样本数据缺乏、影响因素复杂的问题,利用Gram-Schmidt回归方法提取无人机研制费用预测的关键参数,给出了基于Gram-Schmidt回归的军用无人机研制费用预测模型和算法步骤。利用样本数据进行检验,并与逐步多元回归、人工神经网络和偏最小二乘等预测方法进行了比较。结果表明Gram-Schmidt回归方法用于军用无人机研制费用预测具有实用性和优越性。
In this paper,cost prediction of developing a military uninhabited air vehicle（ UAV） is addressed with the fact that the sample data are scarce. To overcome the difficulty resulting from the scarce sample data,Gram-Schmidt regression method is utilized to extract the key parameters. Based on the results obtained by Gram-Schmidt regression,the model and algorithm for predicting military UAV development cost are presented. By exploiting the limited sample data,an instantiated prediction model is built and tested. The proposed method is compared with Sequential Multivariable Regression,Artificial Neural Network,and Partial Least Squares. Results demonstrate the practicability and superiority of applying Gram-Schmidt regression method to the prediction of developing military UAV cost.
出处
《工业工程》
北大核心
2013年第6期29-33,共5页
Industrial Engineering Journal
