CNN-LSTM: A Deep Learning Model for Long-Term Prediction of Tap Water Quality

In recent years, the safety of tap water quality has attracted wide public attention, and it is urgent to improve the tap water quality monitoring mechanism. In this study, water quality data from 33 large water purification plants in Korea over eight years (2014–2022) were used to train the model. To address the problem of instability and nonlinearity of water quality data, we constructed a Convolutional Neural Network-Long short time memory (CNN-LSTM) model for water quality monitoring. On the basis of Long short time memory (LSTM) model, we creatively added Convolutional Neural Network (CNN) convolution module to capture and extract the characteristics of water quality index data. Experimental results show that our model outperforms existing mature models such as LSTM in the prediction accuracy of water quality indicators in time series, and especially in response to nonlinear changes in water quality indicators.