Water quality classification model with small features and class imbalance based on fuzzy rough sets

Water quality has witnessed a significant decline in recent decades due to pollution and various other challenges. Consequently, there is a pressing need for accurate models capable of predicting water quality. This paper proposed a robust fuzzy rough set model for feature selection which can effectively improve classification performance while discarding the irrelevant features. To address the issue of imbalanced data, the synthetic minority oversampling technique (SMOTE) is employed first. Then, a fuzzy rough set is applied to select the most relevant features. Several machine learning models have been performed to determine the most appropriate model applicable based on the performance of the selected feature. The classification was performed with eXtreme gradient boosting (XGBoost), adaptive boosting ensemble (AdaBoost), and random forest. For every combination between the feature selection method and the classifiers, the most appropriate feature subsets were found and used for model training within the training set of the classifiers. Experimental results demonstrate that the proposed model achieves a remarkable accuracy of 98.83%, with precision, sensitivity, specificity, and F1-Score reaching 98.30%, 99.31%, 98.39%, and 98.80%, respectively. This underscores the effectiveness of the XGBoost algorithm in accurately classifying water quality. The suggested model exhibits a high degree of accuracy while employing only a few parameters, making it a promising solution for real-time water quality monitoring systems.