Table 3 Comparison of A-GHN scale outputs with Ground Truth communities. Mutual information between the communities detected in various scales of A-GHN model. Comparison is done by computing the mutual information (MI) between the four communities across different scales. Overall, A-GHN model has higher mutual information with Ground Truth.
From: Attention-based fusion of multiple graphheat networks for structural to functional brain mapping
Model1 vs. Model2 | Mutual Information |
|---|---|
Actual vs. Actual | 1.706 |
Actual vs. A-GHN Scale-0.6 | 0.841 |
Actual vs. A-GHN Scale-0.8 | 1.189 |
Actual vs. A-GHN Scale-1 | 1.312 |
Actual vs. A-GHN Scale-2 | 0.971 |
Actual vs. A-GHN Scale-4 | 0.589 |
Actual vs. A-GHN Scale-6 | 0.546 |
Actual vs. A-GHN Scale-8 | 0.443 |
