CNN-LSTM-attention water quality prediction hybrid model

Zhenhua Mi; Qi Li; Yulong Sha; Zhaoming Wu

doi:10.1117/12.2684594

24 October 2023 CNN-LSTM-attention water quality prediction hybrid model

Zhenhua Mi, Qi Li, Yulong Sha, Zhaoming Wu

Proceedings Volume 12804, Second International Conference on Sustainable Technology and Management (ICSTM 2023); 1280405 (2023) https://doi.org/10.1117/12.2684594
Event: 2nd International Conference on Sustainable Technology and Management (ICSTM2023), 2023, Dongguan, China

Abstract

Water quality prediction is a fundamental task to mitigate water resource crisis and promote water ecological protection and restoration. The traditional LSTM water quality prediction model has the problems of weak generalization performance and low prediction accuracy. In this paper, we take the water quality parameters of Fairy Lake in Jiangxi Province as the research object and propose a hybrid CNN-LSTM-Attention water quality prediction model combining CNN, LSTM and attention mechanism. The model first extracts short-term abstract features of water quality data by CNN, then selects LSTM model to capture long-term dependencies between variables, and finally uses Attention mechanism to determine the importance of different temporal features and assign weights, output prediction results. After experimental validation, the CNN-LSTM-Attention water quality prediction hybrid model has smaller prediction error, and the RMSE and MAE evaluation indexes are better than LSTM prediction model and CNN-LSTM prediction mode.

Citation Download Citation

Zhenhua Mi, Qi Li, Yulong Sha, and Zhaoming Wu "CNN-LSTM-attention water quality prediction hybrid model", Proc. SPIE 12804, Second International Conference on Sustainable Technology and Management (ICSTM 2023), 1280405 (24 October 2023); https://doi.org/10.1117/12.2684594

ACCESS THE FULL ARTICLE

INSTITUTIONAL
Select your institution to access the SPIE Digital Library.

SELECT YOUR INSTITUTION

PERSONAL
Sign in with your SPIE account to access your personal subscriptions or to use specific features such as save to my library, sign up for alerts, save searches, etc.

PERSONAL SIGN IN

No SPIE Account? Create one

PURCHASE THIS CONTENT

SUBSCRIBE TO DIGITAL LIBRARY

50 downloads per 1-year subscription

Members: $195

Non-members: $335 ADD TO CART

25 downloads per 1 - year subscription

Members: $145

Non-members: $250 ADD TO CART

PURCHASE SINGLE ARTICLE

Includes PDF, HTML & Video, when available