Table 4 Influence of different hops on the prediction model
From: gGATLDA: lncRNA-disease association prediction based on graph-level graph attention network
Dataset1 | Dataset2 | |||||
|---|---|---|---|---|---|---|
hop = 1 | hop = 2 | hop = 3 | hop = 1 | hop = 2 | hop = 3 | |
AUC | 0.948 | 0.943 | 0.945 | 0.986 | 0.982 | 0.953 |
Precision | 0.731 | 0.794 | 0.754 | 0.658 | 0.698 | 0.730 |
Recall | 0.965 | 0.900 | 0.926 | 0.999 | 0.995 | 0.988 |
AUPR | 0.953 | 0.948 | 0.951 | 0.985 | 0.983 | 0.950 |
Accuracy | 0.799 | 0.825 | 0.800 | 0.732 | 0.777 | 0.802 |
F1-Score | 0.830 | 0.838 | 0.824 | 0.791 | 0.819 | 0.837 |