Machine Learning Based Water Quality Assessment Under Changing Climatic Conditions
Abstract
Water quality assessment is done to verify water quality in order to maintain ecological balance and protect public health. Though traditional methods are accurate they are often time-consuming and costly. This paper aims to overcome these challenges by using machine learning techniques to efficiently predict water quality levels. Physicochemical features like pH, Dissolved Oxygen (DO), Biological Oxygen Demand (BOD), Nitrate, Conductivity, Fecal Coliform, and Total Coliform were used. A climatic factor, air temperature was also examined to understand how it affects the Water Quality Index (WQI). Changes in temperature can affect dissolved oxygen levels and microbial activity, which are responsible for water quality variations. Four classification models (Random Forest, XGBoost, Gradient Boosting and AdaBoost), along with four regression models (K-Nearest Neighbour (KNN), Decision Tree (DT), Support Vector Regressor (SVR) and Multi-Layer Perceptron (MLP)), were trained on labelled datasets. Grid Search Cross Validation was used for hyperparameter tuning. The models were evaluated using various performance metrics like accuracy, precision, recall, F1-score, RMSE, and MAE. Air temperature significantly affected the results and improved the performance of the models It also increased model reliability and supported effective, data-driven water resource management.