Table 3.
The mean AUC values of five methods on balanced datasets.
| Methods | Enzyme | IC | GPCR | Nuclear Receptor |
|---|---|---|---|---|
| Cao’s work [14] | 0.979 | 0.987 | 0.951 | 0.924 |
| BGL | 0.904 | 0.851 | 0.899 | 0.843 |
| BLM | 0.976 | 0.973 | 0.955 | 0.881 |
| NetLapRLS | 0.956 | 0.947 | 0.931 | 0.856 |
| RLS | 0.978 | 0.984 | 0.954 | 0.922 |
| DAWN (our method) | 0.980 | 0.983 | 0.950 | 0.931 |
Results excerpted from [14]. The best results in each column are in bold faces. BGL: bipartite graph learning; BLM: bipartite local model; NetLapRLS: Laplacian regularized least square based on interaction network; RLS: regularized least square. DAWN: prediction of Drug–tArget interactions based on multi-scale discrete Wavelet transform and Network features.