SARIMA-GRU组合模型的水位预测
Water level forecasting based on SARIMA-GRU combination model
曹寒问 1陈九江 1李小玲1
作者信息
- 1. 南昌工程学院理学院,江西南昌 330099;南昌工程学院工程数学与先进计算重点实验室,江西南昌 330099
- 折叠
摘要
相较于传统的单一模型,组合模型在一定条件下具有更优的预测精度.为验证组合模型是否有利于提高模型的预测精度,本文以长江中游支流澧水石龟山水电站的水位数据为基础,建立SARIMA模型和GRU神经网络模型,并将这两种模型基于方差倒数法和IOWA算子进行组合,最后比较单一模型和组合模型在该水位数据集上的预测精度差异.结果表明,适当的组合方式有利于提高模型预测精度,基于IOWA算子的组合模型具优良的预测性能.
Abstract
Compared with traditional single models,combination models have better predictive accuracy under certain condi-tions.To verify whether the combination model is conducive to improving the prediction accuracy of the model,the water lev-el data of Shigui Mountain hydropower station,a tributary of the Yangtze River in the middle reaches are used as the basis.SARIMA model and GRU neural network model are established.The two models are then combined using the inverse vari-ance weighted average method and the IOWA operator.Finally,the predictive accuracy difference between the single model and the combination model is compared with the water level dataset.The results show that appropriate combination methods are conducive to improving model predictive accuracy,and the combination model based on the IOWA operator has excellent predictive performance.
关键词
SARIMA/GRU神经网络/水位预测/组合模型Key words
SARIMA/GRU neural network/water level forecasting/combination model引用本文复制引用
基金项目
江西省高校人文社会科学研究项目(TJ23101)
出版年
2024
国家科技期刊平台
NSTL
万方数据