基于神经网络的民航客运量的预测研究

Research on predicting civil aviation passenger transport volume based on neural networks

为准确预测民航客运量,解决传统模型无法精准捕捉不稳定客运量的波动问题,选取更可靠的BP模型,以此挖取数据的非线性以及非平稳特征和规律。针对近17年的时间序列民航客运量进行预测研究,构建BP神经网络预测模型,并与传统模型ARIMA对比分析。进行对比预测曲线图,直观地反映出BP模型比ARIMA模型在波动点的预测上更加稳定,表现更好。结果表明,BP模型相比于ARIMA模型有更高的R2和更低的MSE,能够更有效提高民航客运量的预测精度和预测稳定性,为制定航空运输生产计划和发展航空运输业提供了重要参考。

In order to accurately predict civil aviation passenger traffic and solve the problem that traditional models cannot ac⁃curately capture the fluctuation of unstable passenger traffic,a more reliable BP model is selected as a way to mine the nonlinear and nonsmooth features and laws of the data.Aiming at the time series of civil aviation passenger traffic in the past 17 years,we carry out forecasting research,construct the BP neural network forecasting model,and compare and analyze it with the traditional model ARIMA.Comparison of prediction graphs is carried out,which intuitively reflects that the BP model is more stable and per⁃forms better than the ARIMA model in the prediction of fluctuation points.The results show that the BP model has higher R2 and lower MSE than the ARIMA model,which can more effectively improve the prediction accuracy and prediction stability of civil aviation passenger traffic,and provides an important reference for the development of air transportation production plan and the de⁃velopment of air transportation industry.