Data augmentation and machine learning code in the prediction of Mg-based hydrogen storage alloy

Published: 17 November 2025| Version 3 | DOI: 10.17632/gx29c22twj.3
Contributors:
Yichuan Wang, Shuhan Liu

Description

This dataset includes augmented data related to the hydrogen storage capacity of Mg-based hydrogen storage alloys, as well as code used for data augmentation, model building, and model interpretation. The data augmentation code contains physical constraint augmentation, Gaussian noise augmentation, interpolation augmentation, and SMOTE oversampling augmentation. The model implementation includes code for SVR, MLP, RF, XGBoost, and the SHAP method.

Files

Steps to reproduce

We performed both data augmentation and model development using the software R Studio.

Institutions

  • Hebei University of Technology

Categories

Energy Materials, Interpretable Machine Learning

Licence