Skip to main content

Table 6 Statistics about the comparative performance of GEST and our label similarity-incorporated kNN classifiers, measured in terms of the number of classes for which AUC scores are improved by the latter over the former, and the average and maximum improvement in AUC scores.

From: Incorporating functional inter-relationships into protein function prediction algorithms

Dataset Total # classes # Classes improved Average improvement over all classes Maximum improvement
Mnaimneh 137 87 0.0116 (1.86%) 0.1788 (37.83%)
Rosetta 137 65 0.0004 (0.06%) 0.1854 (36.6%)
Krogan 108 60 0.0059 (0.84%) 0.1307 (23.53%)
Combined 136 75 0.0081 (1.04%) 0.2117 (48.85%)