基于判别分析—SVR的民航客运量预测模型研究及应用

Research and application of civil aviation passenger volume model based on discriminant analysis-SVR

为了提高预测民航客运量的能力,考虑到民航客运量与其影响因素之间存在关联,并利用训练样本与测试样本间的马氏距离对惩罚因子进行加权,改进传统的ε支持向量回归机(SVR),构造了基于进化ε-SVR的"影响因素-民航客运量"预测模型.在选择适当的参数和核函数的基础上,对中国民航客运量进行仿真实验,与标准的ε-SVR方法、BP人工神经网络和线性回归方法进行了对比,发现该方法能获得较小的训练相对误差和测试相对误差.

To improve the forecast ability of the passenger traffic volume of civil aviations, take into account the intrinsic relations, between the factors of the impact and regional logistics demand, ＂the factors of impact the passenger traffic volume of civil aviation＇forecast model based on improved ε support vector regression is developed by using Mahalanobis distance between training and testing samples to get weighted penalty coefficients. By selecting appropriate parameters and kernel function, the proposed approach is used for forecasting the passenger traffic volume of civil aviation in china, compared with the normalε support vector regression, BP artificial neural network and linear regression, experimental results show that the training relative error and testing relative error obtained are lower than those by the normalε support vector regression BP artificial neural network and linear regression.