Table 4.
Mean training and validation scores for the best performing machine learning models using the 1.5 times mean liver SUV thresholding segmentation technique.
| Machine Learning Model | Hyperparameters | Features | AUC Mean Training | AUC Mean Validation |
|---|---|---|---|---|
| SUVmax/130 | ||||
| Ridge Regression | C: 1 × 10−5, penalty: l2, solver: saga | Stage Four, Age, PET original GLDM Large Dependence Low Grey Level Emphasis, PET original GLSZM Large Area High Grey Level Emphasis | 0.74 (0.03) | 0.71 (0.09) |
| Support Vector Machine | C: 1, gamma: 0.43727367418726576, kernel: rbf | PET square 10th Percentile, PET square first order Energy | 0.78 (0.02) | 0.73 (0.07) |
| Random Forest | bootstrap: True, max depth: 10, max features: sqrt, min samples leaf: 33, min samples split: 5, n estimators: 90 | Age, PET original shape Elongation, PET original shape Least Axis Length, PET original shape Major Axis Length, PET original shape Maximum 2D Diameter Column, PET original shape Mesh Volume | ||
| SUVmax/64 | ||||
| Ridge Regression | C: 1.0, penalty: l2, solver: liblinear | Stage Three, Age, PET wavelet-LHL GLCM Imc1, PET square GLDM Dependence Variance, PET square GLSZM Small Area Low Grey Level Emphasis | 0.76 (0.02) | 0.73 (0.07) |
| Support Vector Machine | C: 1, gamma: 0.43727367418726576, kernel: rbf | PET square first order 10 Percentile, PET square first order Energy | 0.78 (0.02) | 0.73 (0.07) |
| Random Forest | bootstrap: True, max depth: 10, max features: log2, min samples leaf: 42, min samples split: 6, n estimators: 237 | PET original shape Sphericity, PET original GLSZM Large Area Emphasis | 0.70 (0.02) | 0.69 (0.07) |
l2 = Ridge regression penalty, liblinear = A library for large linear classification, GLSZM = grey level size zone matrix, GLDM = grey level dependence matrix, lbp-3D-m1 = local binary pattern filtered image at level 1, lbp-3D-k = local binary pattern kurtosis image, GLCM = grey level co-occurrence matrix, rbf = radial basis function.