Prediction of Occurrence-State Characterization Parameters of TiO2 Nanoparticles in Aquatic Environments Using Interpretable Machine Learning

Dataset for "Prediction of Occurrence-State Characterization Parameters of TiO₂ Nanoparticles in Aquatic Environments Using Interpretable Machine Learning" This dataset accompanies the above paper and contains experimentally reported hydrodynamic diameter and zeta potential data of TiO₂ nanoparticles in aqueous media, compiled through a systematic literature search of the Web of Science Core Collection, the Nano E-Tox database, and the US EPA ECOTOX knowledge base (71 papers, 1980–2026). The data were used to develop interpretable machine learning models (random forest, gradient boosting, and support vector regression) for predicting the aqueous occurrence-state characterization parameters of TiO₂ nanoparticles.