Table 2.
Abbreviation | Description | Input | Classification Method |
---|---|---|---|
3D-ResNet | 3D Residual Network | 3D preprocessed image | 3D- ResNet |
XGB-HC | XGBoostmodelon hippocampus regions | Flattened 3D hippocampus region extracted from 3D preprocessed image | XGBoost |
XGB | XGBoostmodelon complete image | Flattened 3D hippocampus extended region extracted from 3D preprocessed image | XGBoost |
RTE-XGB | XGBoost model with input dimension reduction | Flattened 3D hippocampus extended region extracted from 3D preprocessed image | RTE XGBoost |
ENS-1 | Ensemble model of all XGBoost models |
Predicted probability of XGB-HC, XGB, RTE-XGB |
XGBoost |
ENS-2 | ENS-1 + 3D-ResNet | Predicted probability of XGB-HC, XGB, RTE-XGB, 3D-ResNet | XGBoost |
ENS-3 | ENS-1 + 3D-ResNet + demographics features | Predicted probability of XGB-HC, XGB, RTE-XGB, 3D-ResNet, demographic features (age and gender) | XGBoost |
ENS-ALL | ENS-1 + 3D-ResNet + demographics features + cognitive scores | Predicted probability of XGB-HC, XGB, RTE-XGB, 3D-ResNet, demographic features (age and gender), cognitive test scores (MMSE and CDR) | XGBoost |