Machine Learning Prediction of Stability and Formation Energy of HCP Non-metallic Impurities in the Presence of Alloying Elements
Venue: 2024 China Materials Conference - Computational Simulation Session, Guangzhou, China
Presentation Type: Oral presentation
Abstract
This presentation discussed the application of machine learning methods to predict the stability and formation energy of non-metallic impurities in HCP metals when alloying elements are present. The study utilized SISSO and SHAP algorithms to develop interpretable prediction models based on extensive DFT calculations.
Key Topics
- Machine learning in computational materials science
- SISSO (Sure Independence Screening and Sparsifying Operator) algorithm
- SHAP (SHapley Additive exPlanations) for feature interpretation
- HCP metal interstitial behavior
- Effects of alloying elements on non-metallic impurities
Research Impact
This work provides a data-driven approach to understanding and predicting interstitial behavior in HCP alloys, which is crucial for designing advanced structural materials with improved mechanical and corrosion properties.