Water-Quality Data Imputation with a High Percentage of Missing Values: A Machine Learning Approach

The monitoring of surface-water quality followed by water-quality modeling and analysis are essential for generating effective strategies in surface-water-resource management. However, worldwide, particularly developing countries, studies limited due to the lack a complete reliable dataset surface-water-quality variables. In this context, several statistical machine-learning models were assessed imputing data at six stations located Santa Lucía Chico river (Uruguay), mixed lotic lentic system. challenge study is represented high percentage missing (between 50% 70%) temporal spatial variability that characterizes competing algorithms implement univariate multivariate imputation methods (inverse distance weighting (IDW), Random Forest Regressor (RFR), Ridge (R), Bayesian (BR), AdaBoost (AB), Hubber (HR), Support Vector (SVR) K-nearest neighbors (KNNR)). According results, more than 76% outcomes considered “satisfactory” (NSE > 0.45). performance shows better results inside reservoir those positioned along mainstream. IDW was model with best RFR, HR SVR. approach proposed expected aid water-resource researchers managers augmenting datasets overcoming issue increase number future related matter.