Predicting Laguna de Bay Water Quality with Water Quality Index and Machine Learning

This study investigates the effectiveness of a Water Quality Index as an attribute in predicting individual water quality parameter values using Multivariate Linear Regression and classifying overall water quality using a Support Vector Machine. We employ a dataset of seven hundred observations from the Laguna Lake Development Authority containing water quality parameters such as Biochemical Oxygen Demand, Dissolved Oxygen, Fecal Coliform, pH, Ammonia, Nitrate, and Inorganic Phosphate. Our findings indicate that incorporating a WQI did not significantly improve the performance of either MLR or SVM models in achieving the stated objectives. However, Multiple Linear Regression achieved superior accuracy when employing GridSearch for hyperparameter optimization.