模态分解与混合模型融合的民航客运量预测
Civil aviation passenger volume prediction based on the fusion of modal decomposition and hybrid models
唐甜甜 1徐海文 1刘浩霖 1于飞 1何梦帆1
作者信息
- 1. 中国民用航空飞行学院,四川 简阳 641419
- 折叠
摘要
为解决民航客运量预测时存在的精度低、高维特征缺失等问题,结合中国民航客运量数据的特征,提出一种由互补集合经验模态分解(CEEMD)与支持向量机(SVM)组合的预测模型.通过CEEMD对数据进行分解,有效处理数据中的复杂特征和趋势.通过粒子群算法(PSO)优化SVM模型的参数,确保模型更好地适应数据特征并提供准确的预测结果.构建CEEMD-PSO-SVM组合预测模型应对复杂的客运量数据,提升预测效果.选取 2005-2024年的客运量数据进行建模,并与CEEMD-SVM、EMD-SVM、EEMD-SVM、EMD-PSO-SVM、EEMD-PSO-SVM模型的预测效果进行比较.仿真结果表明:构建的CEEMD-PSO-SVM模型可有效预测中国民航客运量波动特征下的数据变化趋势.
Abstract
Aviation industry plays an important role in the transportation sector.Timely and accurate prediction of civil aviation passenger volume has gained keen academic interest.To address the low accuracy and missing high-dimensional features,we propose a prediction model combining complementary ensemble empirical modal decomposition(CEEMD)and support vector machine(SVM)in combination with the characteristics of Chinese civil aviation passenger volume.First,the data are decomposed by CEEMD to effectively tackle the complex features and trends.Second,the parameters of the SVM model are optimized by Particle Swarm Optimization Algorithm(PSOA)to ensure the model better adapts to the data features and provides accurate prediction results.Then,a combined CEEMD-PSOA-SVM prediction model is built to analyze the complex passenger volume data and improve the prediction results.The passenger volume data from 2005 to 2024 are selected for modeling and compared with the predictions of CEEMD-SVM,EMD-SVM,EEMD-SVM,EMD-PSO-SVM,EEMD-PSOA-SVM models by employing appropriate performance evaluation indexes.Our results show our CEEMD-PSOA-SVM model generates more accurate predictions of China's civil aviation passenger volumes.
关键词
互补集合经验模态分解/粒子群算法/支持向量机/组合预测/民航客运量预测Key words
complementary set empirical mode decomposition/PSOA/SVM/combination prediction/passenger volume prediction引用本文复制引用
出版年
2024
国家科技期刊平台
NSTL
万方数据