Table 2.
Multi-class F1 scores and their standard deviations, of the iterated five-fold cross validation experiments, for predicting the cognitive status of ADNI participants averaged over each time-point.
| Model | F1(AD) | F1 (MCI) | F1 (HC) | F1 (All) |
|---|---|---|---|---|
| Logistic | 0.265±0.0276 | 0.500±0.0353 | 0.313±0.0562 | 0.396±0.0299 |
| RandomForest | 0.325±0.0201 | 0.415±0.0466 | 0.401±0.0308 | 0.386±0.0325 |
| SVM | 0.289±0.0341 | 0.474±0.0450 | 0.363±0.0254 | 0.396±0.0286 |
| KNN | 0.330±0.0415 | 0.472±0.0524 | 0.410±0.0388 | 0.420±0.0332 |
| MLP | 0.312±0.0588 | 0.475±0.0523 | 0.341±0.0737 | 0.400±0.0366 |
| ElasticNet4 | 0.255±0.070 | 0.447±0.0485 | 0.405±0.0655 | 0.390±0.0284 |
| LinearSVM13 | 0.308±0.038 | 0.448±0.0381 | 0.332±0.0364 | 0.378±0.0311 |
| Our method | 0.496±0.0419 | 0.415±0.0222 | 0.477±0.0308 | 0.459±0.0125 |