Cutting Parameter Optimization of Single Point Cutting Tool of CNC Lathe Machine Using Machine Learning Algorithm

In this study, two machine learning models, Random Forest Regressor (RFR) and Decision Tree Regressor (DTR), were applied to predict surface roughness (SR) in machining processes (turning) based on key input parameters: spindle speed (SS), feed rate (FR), and depth of cut (DOC). The performance of both models was evaluated using R² (R-squared), Mean Squared Error (MSE), Root Mean Squared Error (RMSE), and Mean Absolute Error (MAE). The Decision Tree Regressor achieved near-perfect accuracy with an R² score of 0.9999 and minimal error metrics. The Random Forest Regressor also performed exceptionally well, with an R² of 0.9988 and similarly low error values. Visual analyses, including residual and radar plots, confirmed the high accuracy of both models, with the Decision Tree Regressor slightly outperforming the Random Forest model. However, the Random Forest Regressor's ensemble structure provides better generalization and robustness, making it a more reliable model for larger or more complex datasets. This study concludes that both models are highly effective for predicting surface roughness, but the choice between them should depend on the specific trade-offs between accuracy and generalization needed in a given application.