Hybrid Nesterov-accelerated adaptive moment estimation-differential evolution optimization for long short-term memory-based dissolved oxygen prediction in water quality assessment

Abstract

Accurate and dynamic prediction of water quality indicators is increasingly critical due to rising pollution and water resource insecurity, particularly when dealing with high-dimensional, nonlinear time series data. Dissolved oxygen (DO), a key indicator of aquatic ecosystem health and pollution, requires high prediction accuracy for effective environmental management. This study aims to enhance the accuracy and adaptability of DO prediction by addressing the limitations of traditional deep learning methods, such as slow convergence and local optima. We propose a novel hybrid optimization framework that combines Nesterov-accelerated Adaptive Moment Estimation (Nadam) with the differential evolution algorithm. A dual-population cooperation strategy and an information exchange mechanism were incorporated during the training of a long short-term memory (LSTM) network to achieve a dynamic balance between global exploration and local exploitation. This improves the model’s optimization efficiency and generalization. The research utilized a multivariate water quality time series dataset from Kaggle based on official data monitoring. Correlation analysis was conducted to ensure the scientific validity and effectiveness of the selected input variables. Experimental results demonstrated that the proposed method significantly outperforms traditional optimization strategies for DO prediction. Compared to the original Nadam optimizer, it reduced the mean squared prediction error by 47.8%, exhibiting enhanced adaptability and robustness in complex pollution scenarios. This study presents an effective optimization strategy to improve LSTM performance in water quality forecasting, along with a scalable and interpretable intelligent analysis framework. It provides both theoretical and practical support for water quality forecasting, early warning systems, and intelligent environmental monitoring.

Keywords

Water quality management / Dissolved oxygen prediction / Hybrid optimization / Nadam / Differential evolution

Funding

None.

Conflict of interest

The authors declare no conflicts of interest.

References

[1] Yang J, Chen TS, Sun Q, Huang BY, Ren SJ. The impact of Urban development upon the quality of drinking water sources: Evidence from China. IOP Conf Ser Earth Environ Sci. 2020; 508(1):e012072. doi: 10.1088/1755-1315/508/1/012072

[2] Liu Z, Ying J, He C, et al. Scarcity and quality risks for future global Urban water supply. Landsc Ecol. 2024; 39(10):1-17. doi: 10.1007/s10980-024-01832-0

[3] Ali B, Anuska, Mishra A. Effects of dissolved oxygen concentration on freshwater fish: A review. Int J Fish Aquat Stud. 2022; 10(4):113-127. doi: 10.22271/fish.2022.v10.i4b.2693

[4] Liu W, Lin S, Li X, et al. Analysis of dissolved oxygen influencing factors and concentration prediction using input variable selection technique: A hybrid machine learning approach. J Environ Manage. 2024; 357:120777. doi: 10.1016/j.jenvman.2024.120777

[5] Hu Y, Liu C, Wollheim WM. Prediction of riverine daily minimum dissolved oxygen concentrations using hybrid deep learning and routine hydrometeorological data. Sci Total Environ. 2024; 918:170383. doi: 10.1016/j.scitotenv.2024.170383

[6] Doroudi S, Kheyruri Y, Sharafati A, Hameed AS. Enhancing prediction of dissolved oxygen over Santa Margarita River: Long short-term memory incorporated with multi-objective observer-teacher-learner optimization. J Water Process Eng. 2025; 70:106969. doi: 10.1016/j.jwpe.2025.106969

[7] Lindemann B, Müller T, Vietz H, Jazdi N, Weyrich M. A survey on long short-term memory networks for time series prediction. Procedia CIRP. 2021; 99(2):650-655. doi: 10.1016/j.procir.2021.03.088

[8] Hikmat Haji S, Mohsin Abdulazeez A. Comparison of optimization techniques based on gradient descent algorithm: A review. J Archaeol Egypt. 2021; 18(4):2715-2743.

[9] Dozat T.Incorporating nesterov momentum into Adam. ICLR Work. 2016; 1:1-4.

[10] Yu Y, Gao S, Wang Y, Todo Y.Global optimum-based search differential evolution. IEEE/CAA J Autom Sin. 2019; 6(2):379-394. doi: 10.1109/JAS.2019.1911378

[11] Ren X, Xu Z, Chu W, et al. The discharge of chlorinated effluent from wastewater treatment plants enhances dissolved oxygen in the receiving river: From laboratory study to practical application. Water Res. 2025; 273(9):123012. doi: 10.1016/j.watres.2024.123012

[12] Jiang Y, Zhang L, Wang C, et al. Multi-step prediction of dissolved oxygen in fish pond aquaculture using feature reconstruction-based deep neural network. Comput Electron Agric. 2025; 232(14):109997. doi: 10.1016/j.compag.2025.109997

[13] Ji X, Shang X, Dahlgren RA, Zhang M. Prediction of dissolved oxygen concentration in hypoxic river systems using support vector machine: A case study of Wen-Rui Tang River, China. Environ Sci Pollut Res. 2017; 24(19):16062-16076. doi: 10.1007/s11356-017-9243-7

[14] Khan UT, Valeo C. A new fuzzy linear regression approach for dissolved oxygen prediction. Hydrol Sci J. 2015; 60(6):1096-1119. doi: 10.1080/02626667.2014.900558

[15] Wu J, Li Z, Zhu L, Li G, Niu B, Peng F. Optimized BP neural network for dissolved oxygen prediction. IFAC-PapersOnLine. 2018; 51(17):596-601. doi: 10.1016/j.ifacol.2018.08.132

[16] Malek S, Mosleh MA, Syed SM. Dissolved oxygen prediction using support vector machine. Int J Comput Inform Syst Control Eng. 2014; 8(1):46-50.

[17] Tan W, Zhang J, Liu X, et al. dissolved oxygen prediction based on PCA-LSTM. J Phys Conf Ser. 2022; 2337(1):012012. doi: 10.1088/1742-6596/2337/1/012012

[18] Hao Z. Research on prediction algorithm of dissolved oxygen in aquatic products based on improved lstm algorithm. 2024 IEEE 3rd Int Conf Electr Eng Big Data Algorithms(EEBDA). 2024:86-90. doi: 10.1109/EEBDA60612.2024.10485663

[19] Gandh DR, Harigovindan VP, Rasheed Abdul Haq KP, Bhide A. Attention-driven LSTM and GRU deep learning techniques for precise water quality prediction in smart aquaculture. Aquac Int. 2024; 32(6):8455-8478. doi: 10.1007/s10499-024-01574-5

[20] Lee SY, Jeong DY, Choi J, Jo SK, Park DH, Kim JG. LSTM model to predict missing data of dissolved oxygen in land-based aquaculture farm. ETRI J. 2024; 46(6):1047-1060. doi: 10.4218/etrij.2023-0337

[21] Shuai M, Tian H. Long short term memory based on differential evolution in passenger flow forecasting. J Simul. 2019; 7(1):17-21.

[22] Peng L, Liu S, Liu R, Wang L. Effective long short-term memory with differential evolution algorithm for electricity price prediction. Energy. 2018; 162:1301-1314. doi: 10.1016/j.energy.2018.05.052

[23] Storn RM, Price K. Differential evolution - A simple and efficient heuristic for global optimization over continuous spaces. Australas Plant Pathol. 2009; 38(3):284-287.

[24] Deng W, Shang S, Cai X, Zhao H, Song Y, Xu J. An improved differential evolution algorithm and its application in optimization problem. Soft Comput. 2021; 25:5277-5298. doi: 10.1007/s00500-020-05527-x

[25] Nguyen QV, Miller N, Arness D, Huang W, Huang ML, Simoff S. Evaluation on interactive visualization data with scatterplots. Vis Informatics. 2020; 4(4):1-10. doi: 10.1016/j.visinf.2020.09.004

[26] Lu Z. Comparison of stock price prediction models for linear models, random forest and LSTM. Appl Comput Eng. 2024; 54(1):226-233. doi: 10.54254/2755-2721/54/20241598

[27] Gürsoy Mİ, Alkan A. Investigation of diabetes data with permutation feature importance based deep learning methods. Karadeniz Fen Bilim Derg. 2022; 12(2):916-930. doi: 10.31466/kfbd.1174591

[28] De Smedt T, Daelemans W. Pattern for python. J Mach Learn Res. 2012; 13(1):2063-2067.

[29] Chicho BT, Sallow AB. A comprehensive survey of deep learning models based on keras framework. J Soft Comput Data Min. 2021; 2(2):49-62. doi: 10.30880/jscdm.2021.02.02.005

[30] Karim F, Majumdar S, Darabi H. Insights into lstm fully convolutional networks for time series classification. IEEE Access. 2019; 7:67718-67725. doi: 10.1109/ACCESS.2019.2916828

[31] Liao L, Li H, Shang W, Ma L. An empirical study of the impact of hyperparameter tuning and model optimization on the performance properties of deep neural networks. ACM Trans Softw Eng Methodol. 2022; 31(3):1-40. doi: 10.1145/3506695

[32] Hodson TO, Over TM, Foks SS.Mean squared error, deconstructed. J Adv Model Earth Syst. 2021; 13(12):e2021MS002681. doi: 10.1029/2021MS002681