Fig. 9

Comparison of MSE, RMSE, and MAE values for various regression models trained and tested on the scaled dataset using StandardScaler. The data includes six input variables (X1–X6), and the target is ‘Removal (%)’. An 80/20 train-test split was applied. The RANSACRegressor’s MSE value (85.89) was used as a benchmark for comparison. Each model was evaluated on the test set to assess generalization performance.