Dataset_PB_310812/2022-4

This record contains the dataset, prediction outputs, model-performance metrics, and supporting files associated with the manuscript “Machine-learning prediction of pH and dissolved oxygen in Pontal Bay, northeastern Brazil: implications for water-quality monitoring”. The dataset includes physicochemical variables, spatial descriptors, and hydroclimatic predictors used to train and evaluate ANFIS and MLP models for pH and dissolved oxygen prediction in Pontal Bay, Ilhéus, Bahia, Brazil. The final modelling dataset comprised 56 complete observations obtained after screening an initial set of 65 sampling records.