Research Article
MR-Based Radiomics for Differential Diagnosis between Cystic Pituitary Adenoma and Rathke Cleft Cyst
Table 2
The mean AUC value of fivefold crossvalidation using different combinations of feature selection and classifiers in the training set.
| Model | Classifier | ANN | SVM | AdaBoost | RF |
| T1WI model | 0.722 | 0.756 | 0.682 | 0.756 | T2WI model | 0.847 | 0.835 | 0.779 | 0.817 | Postcontrast T1WI model | 0.867 | 0.850 | 0.829 | 0.847 | Multiparametric model | 0.890 | 0.889 | 0.845 | 0.868 | Semantic model | 0.902 | 0.842 | 0.844 | 0.873 | Combined radiomics and semantic model | 0.924 | 0.907 | 0.849 | 0.889 |
|
|