Table 1 Statistics of performance in cross validation of the non-redundant dataset from PDB

From: A deep learning framework to predict binding preference of RNA constituents on protein surface

Metrics NonSite Phosphate Ribose Adenine Guanine Uracil Cytosine
AUROC 0.97 0.93 0.84 0.67 0.67 0.65 0.66
F1-score (macro) 0.90 0.70 0.63 0.47 0.38 0.48 0.32
Recall (macro) 0.88 0.82 0.63 0.46 0.38 0.45 0.37
Precision (macro) 0.92 0.61 0.63 0.48 0.38 0.51 0.29
F1-score (micro) 0.90 0.70 0.64 0.47 0.41 0.48 0.32
Recall (micro) 0.88 0.81 0.64 0.47 0.40 0.46 0.36
Precision (micro) 0.92 0.61 0.64 0.48 0.41 0.51 0.29