Table 7.
Confusion matrix (in percentage) for ADNet-DA, provided by CADDementia.
Prediction | ||||
---|---|---|---|---|
CN | MCI | AD | ||
Actual | CN | 68.2 | 25.6 | 6.2 |
MCI | 51.6 | 37.7 | 10.7 | |
AD | 29.1 | 21.4 | 49.5 |
Values are adjusted relative to the actual class, i.e., divided by the row sum. This way, the main diagonal represents the true positive fraction (TPF) for each class. AD, Alzheimer's disease; CN, cognitively normal; MCI, mild cognitive impairment.