Featurized Materials Dataset for Formation Energy Prediction (derived from Materials Project)

This dataset contains featurized materials data derived from the Materials Project database. The data has been processed and transformed into machine learning–ready features for the task of predicting formation energy. The dataset was used in training and evaluating deep learning models with Bayesian optimization, k-fold cross-validation, and early stopping. It provides a reproducible foundation for further research in computational materials science and machine learning applications in predicting materials properties. Original raw data was obtained from the Materials Project and further featurized by the author.