Table 2 Area under ROC curve (AUC, in %) for Naive Bayes, for different datasets and different values of the GO term occurrence threshold (t)
From: A data mining approach for classifying DNA repair genes into ageing-related or non-ageing-related
Dataset Id | PPI-related attributes | t = 3 (301 GO terms) | t = 7 (157 GO terms) | t = 11 (101 GO terms) |
|---|---|---|---|---|
D1 | none | 75.9 | 74.9 | 71.9 |
D2 | #partners | 76.0 | 75.3 | 74.0 |
D3 | #partners + 10 BPI attr's | 78.3 | 77.1 | 76.6 |
D4 | #partners + 20 BPI attr's | 80.5 | 80.1 | 79.4 |
D5 | #partners + 30 BPI attr's | 80.7 | 80.2 | 82.6 |