Optimized Deep Learning Model for Water Quality Prediction

The quality of water exerted great influence on man, animals, plants, industries, and the environment. The past decades have seen a worsening of pollution and contamination of water sources. A crucial measure of water purity, the Water Quality Classification and Water Quality Index are being predicted in this investigation. In order to speed up water quality verification, scientists and the government are in need of new tools to assist them. Using feature extraction techniques, a regression model is analyzed to identify which features accurately detect water quality. The authors have used deep neural networks to predict WQI and WQC with optimization and tuning techniques to maximize prediction accuracy. For predicting water quality, a novel Bi-LSTM with Dynamic Particle Swarm Optimization (DPSO) model has been proposed. A water quality prediction dataset has been used to test the proposed system. The dataset used to predict the WQI contained 2000 cases and 7 features. Optimization-based search techniques were necessary for fine-tuning and optimizing parameters for regression models. Model performance was evaluated using error analysis metrics. An accuracy rate of 99.3 % is achieved by this model when predicting WQI.