Table 2.
Results of six methods on two classification tasks. ACC: ACCuracy; SEN: SENsitivity; SPE: SPEcificity; The number in the bracket (i.e., “(No.)”) denotes the average number of features involved in the course of classification.
| Method | lMCI vs. eMCI (%)
|
eMCI vs. NC (%)
|
||||||
|---|---|---|---|---|---|---|---|---|
| ACC (No.) | SEN | SPE | AUC | ACC (No.) | SEN | SPE | AUC | |
| Baseline | 56.6 (18.6) | 55.8 | 57.1 | 55.1 | 59.4 (26.8) | 66.1 | 52.0 | 56.9 |
| TVF | 57.6 (20.8) | 48.8 | 64.3 | 59.9 | 60.4 (18.4) | 64.3 | 56.0 | 59.1 |
| SVF | 63.6 (14.3) | 60.5 | 66.1 | 65.8 | 63.2 (20.3) | 71.4 | 54.0 | 64.4 |
| CV2F | 64.6 (22.2) | 53.5 | 73.2 | 72.6 | 68.9 (41.0) | 66.1 | 72.0 | 70.5 |
| gLASSO | 74.7 (24.4) | 65.1 | 82.1 | 72.9 | 73.6 (44.2) | 69.6 | 78.0 | 71.2 |
| Proposed | 78.8 (29.2) | 74.4 | 82.1 | 78.3 | 78.3 (46.6) | 82.1 | 74.0 | 77.1 |