摘要
对飞机噪声距离特性(NPD)回归模型的选用进行了研究。通过NPD曲线特征分析给出了线性模型、多项式模型、神经网络和支持向量机(SVM)等四种备选回归模型,结合F16战斗机NPD数据实例对比研究了各种模型的回归与预测效果。结果表明,SVM模型误差最小,性能最稳定,其它模型回归内插精度较高,但外推预测误差较大,性能不够稳定。
On the selection of aircraft noise-distance characteristic （NPD） regression model, By analyzing the characteristic of NPD curves, four regression models such as linear model, polynomial model, neural network and support vector machine （SVM） are nominated, and their effect are studied and compared with an example of F16 fighter plane＇s NPD data. The results indicate that SVM model is the stablest and accuratest while other models work well in interpolation but not in extrapolation.
出处
《科学技术与工程》
北大核心
2013年第15期4470-4474,共5页
Science Technology and Engineering
基金
空军后勤部科研项目(BKJ10J016)资助
关键词
飞机噪声距离特性
回归模型
外推预测
模型优选
aircraft noise-distance characteristic （NPD） model optimization regression model extrapolative prediction
