Predicting and Enhancing Ductility in X₂YZ-type Heusler Alloys Using Machine Learning and Abinitio Calculations

X<inf xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</inf>YZ-type Heusler alloys, such as Ni<inf xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</inf>MnGa, exhibit promising properties like magnetic shape memory, thermoelectric effects, and superconductivity, making them ideal for smart sensors and refrigeration. However, their inherent brittleness limits their mechanical performance, especially in applications requiring elastic or shape memory effects. This study utilizes machine learning (ML) to predict the ductility of these alloys, applying a random forest model combined with advanced feature selection techniques. By training the model on a dataset Heusler alloys from Materials-Project, the study identifies key factors influencing ductility. First-principles calculations validate the predictions, revealing that alloys like Pt<inf xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</inf>TaAl show exceptional ductility, surpassing other known Heusler alloys.