Surface Roughness Prediction in Ball End Milling of AA6061 by Various Machine Learning Algorithms

This work uses machine learning (ML) models to estimate the roughness of surface texture in the manufacture of aluminum by ball end milling operation. The ball end tool, inclination angle, feed rate along the Y-axis, feed rate along the X-axis, spindle speed, and depth of cut are the production process input parameters. Four ML methods were used in this study to estimate the roughness of surface texture: Linear Regression (LIN), Extreme Gradient Boosting (XGB), Gradient Boosting Regression (GBR), Support Vector Regression (SVR) and Random Forest Regression (RFR) (Ra). Different error metrics, such as root mean squared error (RMSE) and mean absolute percentage error (MAPE), were used to investigate the effectiveness of ML models. The GridSearchCV algorithm was used to find the optimum hyperparameters in order to improve the predictive effect of each model. The results exhibited that the XGB model performed the best (exhibiting the lowest RMSE and MAPE using both training and testing datasets).