Machine Learning Enabled Discovery of Photocatalytic ABX3 Perovskite Materials

This repository contains the dataset and code for machine learning-based prediction of band edge positions (conduction band minimum, E_CBM, and valence band maximum, E_VBM) in ABX₃ perovskite materials for photocatalytic applications environmental remediation. Dataset Compounds: 209 ABX₃ perovskite materials Features: 15 elemental descriptors derived from A-site, B-site, and X-site constituents Targets: Conduction band minimum (E_CBM) Valence band maximum (E_VBM) Ground truth: DFT-calculated values Methods Six machine learning algorithms were benchmarked: CatBoost XGBoost Random Forest LightGBM Support Vector Machine (SVM) Decision Tree Model interpretability was achieved through SHAP (SHapley Additive exPlanations) analysis to extract physical insights linking feature importance to electronic structure principles.