AI-Powered Water Quality Index Prediction: Unveiling Machine Learning Precision in Hyper-Arid Regions

Water is a vital resource essential for all life, and its quality has been compromised by pollution and contamination in recent decades. The Water Quality Index (WQI) is crucial for evaluating water validity for several purposes, including drinking. Accurate WQI prediction allows for proactive strategies to combat water contamination, preserve public well-being, and guarantee access to safe water sources. This study introduces a novel approach utilizing advanced Machine Learning (ML) techniques for WQI prediction, demonstrating substantial improvements over traditional methods. The methods include the Ridge Model, Lasso Model, Random Forest (RF) Model, Extra Trees (ExT) Model, AdaBoost (AB) Model, XGBoost (XGB) Model, Gradient Boosting (GB) Model, LightGBM Model, Linear Regression (LR) Model, K-nearest neighbor (KNN) Model, Regressor (R) Model, Decision Tree (DT) Model, Multi-layer Perceptron (MLP) Model and Support Vector Regressor (SVR) Model, to determine the most effective models for predicting WQI. The proposed models are trained on a publicly available dataset from 145 groundwater well samples collected between January and April 2018 in Abu Dhabi, the United Arab Emirates (UAE). The models' performance was assessed using various metrics, including Mean Square Error (MSE), Mean Absolute Error (MAE), Root Mean Square Error (RMSE), Adjusted R-squared, Mean Absolute Percentage Error (MAPE) and R-squared (R2). Experimental results indicate promising performance across all models. In particular, the LR Model proved to be exceptionally accurate, precisely predicting WQI values with 100% accuracy during testing. According to the experimental findings, this model surpassed others in regression tasks, achieving an R2 value of 100% in WQI prediction. The proposed research confirms the effectiveness of ML algorithms in the field of Water Resources and will serve as a reference for the researchers working in the field of WQI prediction.