Estimation of Water quality Index using Machine Learning Algorithms for Indus River Khairabad
This study evaluated the effectiveness of machine learning algorithms in estimating the Water Quality Index (WQI) for surface water from the Indus River at Khairabad, Pakistan. Data collected from the Water and Power Development Authority (WAPDA) were analyzed using Support Vector Regression (SVR), Random Forest, AdaBoost, Decision Tree, and K-Nearest Neighbors (KNN) to enhance water quality assessment. Among the models tested, AdaBoost exhibited the highest performance, achieving an R² score of 0.99 on the training set and 0.90 on the testing set, demonstrating its superior predictive capabilities. The study highlighted the advantages of integrating machine learning into water quality monitoring, emphasizing automation, efficiency, and accuracy. The findings underscored the potential of these techniques to facilitate real-time monitoring, optimize resource management, and contribute to sustainable water quality maintenance. However, challenges such as continuous data collection, model updates, and the need for skilled personnel were identified. The study recommended integrating machine learning models, particularly AdaBoost, into regular monitoring systems, expanding datasets, and fostering collaborations between research institutions and environmental agencies to enhance predictive accuracy and decision-making in water resource management.
Keywords: Water Quality Index (WQI), Machine Learning, AdaBoost, Random Forest, Support Vector Regression (SVR), Surface Water, Environmental Monitoring
