A machine learning approach to predict coefficient of friction of laser cladding high-entropy alloy coatings

Traditional method for evaluating the coefficient of friction (COF) of laser cladding high-entropy alloy (HEA) coatings, relying on extensive experiments, is costly and inefficient. To address these limitations, this study proposes an efficient machine learning (ML) approach. Random Forest, Gradient Boosting Decision Tree, and XGBoost models were developed as base models to predict the COF of HEA coatings. Feature engineering and a voting strategy were employed to enhance model accuracy and generalization ability, yielding substantial improvements in performance. The optimal model achieved a testing R 2 of 0.97. A satisfactory agreement was achieved between the predicted COFs and the experimental measurements as well as their trend. Additionally, the impact of each feature on the COF was clarified using SHAP (Shapley Additive Explanations). The results indicate that this ML approach is a promising tool for predicting the COF of HEA coatings and can serve as a valuable aid in the design and optimization of coatings.