基于组合模型的月降水量预测研究  

Prediction of Monthly Precipitation Based on Combined Model

作　　者：程桂芳 王雪敏 CHENG Gui-fang;WANG Xue-min(School of Mathematics and Statistics,Zhengzhou University,Zhengzhou 450001,China)

机构地区：[1]郑州大学数学与统计学院,河南郑州450001

出　　处：《水电能源科学》2023年第4期13-16,共4页Water Resources and Power

基　　金：国家自然科学基金项目(11971444);2021年度河南省高等教育教学改革研究与实践重点项目(2021SJGLX060);2022年河南省医学科技攻关联合共建项目(LHGJ20220518);2022年度河南省社科联调研课题项目(SKL-2022-405)。

摘　　要：近年来,由降水量过多或过少引起的灾害日益增加,因此准确地预测降水量对人类的生活和社会的发展具有重大意义和实际应用价值。基于郑州市1990~2019年的月降水量数据,分别利用SARIMA、Prophet、LSTM单一模型对郑州市2020~2021年的月降水量进行预测。为了提高月降水量的预测精度,提出SARIMA-EMD-LSTM、Prophet-EMD-LSTM两种组合模型,实证表明这两种组合模型预测精度更高,均方根误差显著减少,其中Prophet-EMD-LSTM组合模型的预测效果相对较优。最后利用该模型对郑州市2022年4~12月的月降水量进行了预测,精度较高。In recent years,it causes disasters increasingly by too much or too little precipitation.Therefore,accurate prediction of precipitation is of great significance and practical application value to human life and social development.Based on the monthly precipitation data of Zhengzhou from 1990 to 2019,monthly precipitation was forecasted from 2020 to 2021 by utilizing SARIMA,Prophet and LSTM model,respectively.In order to improve the prediction accuracy of the model for monthly precipitation,two combined models of the SARIMA-EMD-LSTM and Prophet-EMD-LSTM were proposed.Empirical analysis shows that the proposed two combined models have higher prediction accuracy and decrease the root mean square error significantly.Furthermore,Prophet-EMD-LSTM model has comparatively better prediction effect.The monthly precipitations in Zhengzhou from April to December,2022 were forecasted with higher precision.

关 键 词：降水量预测 SARIMA PROPHET EMD LSTM 组合模型 

分 类 号：TV124[水利工程—水文学及水资源]