Table 5.
The results of Krogan PPI data and biomedical literature PPI data.
| COACH | CMC | ClusterONE | COAN | |||||
|---|---|---|---|---|---|---|---|---|
| F | σF | F | σF | F | σF | F | σF | |
| Krogan dataset | 0.441 | - | 0.358 | - | 0.401 | - | 0.451 | - |
| Krogan + random I | 0.439 | 0.002 | 0.353 | 0.004 | 0.379 | 0.014 | 0.445 | 0.002 |
| Krogan + random II | 0.436 | 0.004 | 0.349 | 0.006 | 0.354 | 0.021 | 0.449 | 0.006 |
| Krogan + random III | 0.433 | 0.004 | 0.347 | 0.006 | 0.34 | 0.018 | 0.444 | 0.007 |
| Krogan + Aimed | 0.457 | - | 0.411 | - | 0.417 | - | 0.464 | - |
| Krogan + BioInfer | 0.453 | - | 0.366 | - | 0.405 | - | 0.458 | - |
| Krogan + IEPA | 0.444 | - | 0.398 | - | 0.393 | - | 0.453 | - |
| Krogan + HPRD50 | 0.445 | - | 0.384 | - | 0.389 | - | 0.463 | - |
| Krogan + LLL | 0.454 | - | 0.393 | - | 0.404 | - | 0.453 | - |
Krogan + random I, Krogan + random II, and Krogan + random III show the results of randomly adding 1000, 2000 and 3000 PPIs to the Krogan dataset, respectively. The highest F-score of each approach is shown in bold.