Classification results under different settings on ADNI II dataset. The classification task is to classify HC and AD
| Models . | Settings . | Accuracy . | Sensitivity . | Specificity . |
|---|---|---|---|---|
| GCN | Naive | 81.3% | 88.9% | 75.0% |
| Extra input | 80.0% | 83.3% | 77.3% | |
| Hard parameter sharing | 83.8% | 83.3% | 84.1% | |
| Soft parameter sharing | 83.8% | 86.1% | 81.2% | |
| LSTM | Naive | 78.8% | 86.1% | 72.7% |
| Extra input | 80.0% | 80.6% | 79.5% | |
| Hard parameter sharing | 81.3% | 77.8% | 84.1% | |
| Soft parameter sharing | 85.0% | 80.6% | 88.6% | |
| DS-GCN | Naive | 83.8% | 80.6% | 86.4% |
| Extra input | 86.3% | 88.9% | 84.1% | |
| Hard parameter sharing | 87.5% | 91.7% | 81.8% | |
| Soft parameter sharing | 90.0% | 91.7% | 88.6% |
| Models . | Settings . | Accuracy . | Sensitivity . | Specificity . |
|---|---|---|---|---|
| GCN | Naive | 81.3% | 88.9% | 75.0% |
| Extra input | 80.0% | 83.3% | 77.3% | |
| Hard parameter sharing | 83.8% | 83.3% | 84.1% | |
| Soft parameter sharing | 83.8% | 86.1% | 81.2% | |
| LSTM | Naive | 78.8% | 86.1% | 72.7% |
| Extra input | 80.0% | 80.6% | 79.5% | |
| Hard parameter sharing | 81.3% | 77.8% | 84.1% | |
| Soft parameter sharing | 85.0% | 80.6% | 88.6% | |
| DS-GCN | Naive | 83.8% | 80.6% | 86.4% |
| Extra input | 86.3% | 88.9% | 84.1% | |
| Hard parameter sharing | 87.5% | 91.7% | 81.8% | |
| Soft parameter sharing | 90.0% | 91.7% | 88.6% |
Classification results under different settings on ADNI II dataset. The classification task is to classify HC and AD
| Models . | Settings . | Accuracy . | Sensitivity . | Specificity . |
|---|---|---|---|---|
| GCN | Naive | 81.3% | 88.9% | 75.0% |
| Extra input | 80.0% | 83.3% | 77.3% | |
| Hard parameter sharing | 83.8% | 83.3% | 84.1% | |
| Soft parameter sharing | 83.8% | 86.1% | 81.2% | |
| LSTM | Naive | 78.8% | 86.1% | 72.7% |
| Extra input | 80.0% | 80.6% | 79.5% | |
| Hard parameter sharing | 81.3% | 77.8% | 84.1% | |
| Soft parameter sharing | 85.0% | 80.6% | 88.6% | |
| DS-GCN | Naive | 83.8% | 80.6% | 86.4% |
| Extra input | 86.3% | 88.9% | 84.1% | |
| Hard parameter sharing | 87.5% | 91.7% | 81.8% | |
| Soft parameter sharing | 90.0% | 91.7% | 88.6% |
| Models . | Settings . | Accuracy . | Sensitivity . | Specificity . |
|---|---|---|---|---|
| GCN | Naive | 81.3% | 88.9% | 75.0% |
| Extra input | 80.0% | 83.3% | 77.3% | |
| Hard parameter sharing | 83.8% | 83.3% | 84.1% | |
| Soft parameter sharing | 83.8% | 86.1% | 81.2% | |
| LSTM | Naive | 78.8% | 86.1% | 72.7% |
| Extra input | 80.0% | 80.6% | 79.5% | |
| Hard parameter sharing | 81.3% | 77.8% | 84.1% | |
| Soft parameter sharing | 85.0% | 80.6% | 88.6% | |
| DS-GCN | Naive | 83.8% | 80.6% | 86.4% |
| Extra input | 86.3% | 88.9% | 84.1% | |
| Hard parameter sharing | 87.5% | 91.7% | 81.8% | |
| Soft parameter sharing | 90.0% | 91.7% | 88.6% |
This PDF is available to Subscribers Only
View Article Abstract & Purchase OptionsFor full access to this pdf, sign in to an existing account, or purchase an annual subscription.
