This repository contains partial implementation of the paper: "Prediction of thermal conductivity in multi-component magnesium alloys based on machine learning and multiscale computation" (2024).
This project combines machine learning with multiscale computation to predict thermal conductivity in multi-component magnesium alloys. The model achieves:
- 2.16% MAPE for ternary and simpler Mg alloy systems
- 13.60% MAPE for quaternary and higher-order novel systems
This is the official implementation of our thermal conductivity prediction model for multi-component magnesium alloys. We open-source:
- Core model implementation
- Feature engineering pipeline
- Sample dataset
- Evaluation scripts