Table 3.
The Comparison of AUC, AUPRC, 
-value and Params between various models
| Model | Network 
|
Network 
|
Params | ||||
|---|---|---|---|---|---|---|---|
| AUC | AUPRC |
-value |
AUC | AUPRC |
-value |
||
| DTINet | 0.879 | 0.906 | 5.31e−10 | 0.889 | 0.900 | 2.08e−7 | * |
| NeoDTI | 0.955 | 0.889 | 4.84e−8 | 0.946 | 0.846 | 1.99e−4 | 9.98e10 |
| GCN-DTI | 0.918 | 0.897 | 1.76e−8 | 0.922 | 0.914 | 1.61e−4 | * |
| IMCHGAN | 0.956 | 0.903 | 1.06e−4 | 0.946 | 0.929 | 1.35e−3 | 4.44e5 |
| EEG-DTI | 0.954 | 0.964 | 2.24e−5 | 0.942 | 0.941 | 2.41e−4 | 1.84e6 |
| HampDTI | 0.928 | 0.927 | 5.51e−9 | – | – | – | 9.92e4 |
| DeepConv-DTI | 0.909 | 0.917 | 3.23e−8 | – | – | – | 1.52e6 |
| TripletMultiDTI | 0.991 | 0.990 | 6.53e−4 | – | – | – | 1.53e7 |
| AMGDTI | 0.977 | 0.977 | 0.973 | 0.971 | 1.33e5 | ||
‘-’ indicates that the model is not applicable to this dataset. ‘
-value’ is calculated using AUC as a statistic. ‘*’ indicates that the model is not convenient for parameter statistics. ‘Params’ represents the parameter quantity of the model.