7 Appendix

7.1 Tuning SVM Hyperparameters

7.1.1 Without PCA

Support Vector Machines with Radial Basis Function Kernel

9685 samples
   4 predictor
  
No pre-processing
Resampling: Cross-Validated (10 fold)
Summary of sample sizes: 8717, 8715, 8718, 8717, 8715, 8716, ...
Resampling results across tuning parameters:

# Train the SVM model with the best parameters
  C     sigma  RMSE       Rsquared   MAE
   0.1  0.5    0.1489409  0.3791848  0.1152221
   0.1  1.0    0.1496952  0.3715801  0.1160810
   0.1  2.0    0.1518043  0.3528460  0.1183108
   1.0  0.5    0.1492998  0.3810286  0.1149792
   1.0  1.0    0.1508168  0.3700282  0.1161860
   1.0  2.0    0.1532086  0.3523641  0.1184146
  10.0  0.5    0.1515176  0.3683280  0.1165050
  10.0  1.0    0.1563312  0.3389117  0.1205078
  10.0  2.0    0.1634327  0.2997668  0.1269389
   
RMSE was used to select the optimal model using the smallest value.
The final values used for the model were sigma = 0.5 and C = 0.1.

7.1.2 With PCA

7.2 Tuning XGBoost Hyperparameters

7.2.1 Without PCA

7.2.2 With PCA

7.3 Tuning Random Forest Hyperparameters

7.3.1 Without PCA

7.3.2 With PCA