广东水利水电Issue(6)：28-36,9.
基于深度学习的优化寻参模型在河流水位预测中的应用
Application of Optimized Parameter Search Model Based on Deep Learning in River Water Level Prediction
摘要
Abstract
Flood disasters are one of the most serious natural disasters,and the losses caused by floods in my country are huge every year.This study proposes a Long Short-Term Memory(LSTM)composite model based on the Sparrow Search Algorithm(SSA)optimization to improve the accuracy of water level prediction which namely the SSA-LSTM model.The model was applied to the water level prediction experiment of the Memacuo Lake in the Luma River in the western Qinghai-Tibet Plateau,and compared with the traditional LSTM model.MSE(mean square error),RMSE(root mean square error),R2(determination coefficient),MAE(mean absolute error)and other parameters were used as evaluation indexes,and were observed by Taylor diagram.It is found that the performance of the optimized model is better.The experimental results show that the SSA-LSTM composite model has better data fitting degree and higher prediction accuracy than the simple LSTM model.The root mean square error RMSE decreased from 0.392 to 0.335,the error accuracy decreased by 14.54%,and the coefficient of determination R2 increased from 0.777 to 0.835.The experimental data show that the SSA-LSTM combined model with sparrow optimization parameter searching has better fitting accuracy and stability in water level prediction,which provides a new way to carry out the water level prediction of small and medium-sized rivers.关键词
深度学习/长短时记忆网络/麻雀优化算法/水位预测Key words
deep learning/LSTM/SSA/water level prediction分类
计算机与自动化引用本文复制引用
聂影,刘永宏,陈俞强..基于深度学习的优化寻参模型在河流水位预测中的应用[J].广东水利水电,2025,(6):28-36,9.基金项目
广东省攀登计划项目(编号:pdjh2023b1132) （编号:pdjh2023b1132）
清远市科技智库专项项目(项目编号:QYKX2024001) （项目编号:QYKX2024001）
广东省教育厅(重点专项)项目(编号:2022ZDZX1073、2023ZDZX1086). （重点专项）
