Water Quality Assessment Leveraging CNN-LSTM and Gradient Boosting for Prediction and Classification

Accurate prediction of water quality can support in water resource management by the early warning of water pollution. Water quality assessment is crucial for ensuring the availability of safe drinking water and maintaining public health. This study focuses on developing a comprehensive framework for prediction and classification of water quality using advanced machine learning techniques. The Water Quality Index (WQI), which is a numerical representation of water quality, is predicted using hybrid CNN- LSTM model, which effectively shows the nonlinear relationships between water quality parameters such as pH, turbidity, and dissolved oxygen. Additionally, a Gradient Boosting algorithm is employed to categorize the quality of water based on the WQI values whether the water is potable or not. The proposed system is further enhanced with an intuitive user interface, enabling users to input water quality parameters and receive real-time feedback on whether the water is suitable for drinking. The results shows that the proposed models can predict WQI accurately and classify the water quality. Prediction results demonstrated that CNN- LSTM hybrid model performed better for predicting the WQI values and the gradient Boosting algorithm has achieved the accuracy of 97% for the WQI classification. Moreover, CNN- LSTM hybrid model has also achieved the accuracy of 94% for WQI prediction. These kind of researches can significantly contribute to water management.