Application of Machine Learning and Regression Models in the Prediction of Ball Burnishing Parameters for Aluminum Alloy

This study proposes the use of machine learning models to predict the arithmetic mean roughness (Sa) and Vickers hardness (HV) of EN-AW-1350 aluminum alloy after dry ball burnishing. The input parameters considered are ball diameter (mm), number of passes, and depth of cut (mm). Six regression methods Decision Tree (DT), Linear Regression (LR), eXtreme Gradient Boosting (XGBoost), AdaBoost, Support Vector Machine (SVM), and Random Forest (RF) were evaluated using an 80/20 training/test split. XGBoost demonstrated the best predictive performance, achieving the lowest root mean square error (RMSE) and a coefficient of determination (R²) close to 1 for both output responses. The developed models offer an efficient alternative to time-consuming experimental procedures and can effectively guide process optimization