基于灰色GM(1,N)马尔科夫神经网络的民航旅客运输服务投诉预测模型  

Forecast Model of Civil Aviation Passenger Transportation Service Complaint Based on the Grey GM(1,N)-Markov-Neural Network

作　　者：户志浩 赵健 陈洪根 HU Zhihao;ZHAO Jian;CHEN Honggen(School of Management Engineering,Zhengzhou University of Aeronautics,Zhongzhou 450006,China;Luo Yang Polytechnic,Luoyang 471000,China)

机构地区：[1]郑州航空工业管理学院管理工程学院,河南郑州450046 [2]洛阳职业技术学院,河南洛阳471000

出　　处：《郑州航空工业管理学院学报》2025年第2期67-76,共10页Journal of Zhengzhou University of Aeronautics

摘　　要：针对民航旅客服务投诉预防问题,综合使用灰色GM(1,N)预测模型、马尔科夫过程和神经网络方法,提出民航旅客运输服务投诉预测的灰色GM(1,N)马尔科夫神经网络预测模型。该模型包含初步预测、修正、调整三个过程。首先,利用灰色GM(1,N)对历史数据进行预测,得到初步预测结果;然后,针对灰色预测结果往往呈现弱波动无法反映真实数据的强波动性问题,利用马尔科夫模型对预测残差进行修正,得到拟合度更高的预测修正值;在此基础上,针对灰色马尔科夫模型在处理民航旅客服务投诉等复杂非线性数据时存在精度不足问题,进一步引入神经网络模型对预测结果进行调整和优化。实证研究结果表明,相对于仅用灰色预测模型或灰色马尔科夫预测模型,该模型能够较好地捕捉投诉数据的趋势和随机波动性,得到了更加精确的预测结果。To address the issue of preventing service complaints from civil aviation passengers, a Grey GM(1, N)-Markov-Neural Network prediction model for forecasting service complaints in civil aviation passenger transportation is proposed by integrating the Grey GM(1,N) prediction model,Markov process,and neural network methods.This model consists of three processes:preliminary prediction,correction,and adjustment.Firstly,the Grey GM(1,N) model is used to predict historical data,providing an initial forecast.Then,to address the issue that Grey predictions typically exhibit weak fluctuations and fail to reflect the strong volatility of actual data,the Markov model is applied to correct prediction residuals,resulting in more accurate corrected predictions.Building on this,considering the limited accuracy of the Grey-Markov model in handling complex nonlinear data such as civil aviation passenger service complaints,a neural network model is introduced to further adjust and optimize the predictions.Empirical results demonstrate that the proposed model outperforms both the standalone Grey prediction model and the Grey-Markov model in capturing trends and random fluctuations in complaint data, yielding more accurate forecasting results.

关 键 词：民航旅客服务投诉 灰色GM(1 N) 马尔科夫模型 神经网络 

分 类 号：V355[航空宇航科学与技术—人机与环境工程]