A Comprehensive Study on Machine Learning Techniques for Prediction of Material Properties

ABSTRACT Machine learning (ML) is becoming a valuable tool for materials science, driving both basic and applied research. It promises to be especially helpful to model complicated surface dynamics and material property prediction—essential for semiconductor manufacturing, thin‐film deposition, and nanotechnology. Although existing ML applications in materials science tend to be straightforward fitting procedures or small‐scale data, this research investigates its ability in 1+1 and 2+1 dimensional growth models and displays precise predictions for material stability and crystal structures. The paper includes ML basics (algorithms, descriptors, databases) and uses ML for surface dynamics, emphasizing ML's contribution to material property control.