Skip to main content

Table 6 Performance comparison of the six methods on the CASP9 data set.

From: LigandRFs: random forest ensemble to identify ligand-binding residues from sequence information alone

Method

Type

# of targets

Sen (%)

MCC

Prec (%)

F1 (%)

LigandRFs

Random Forest

30

42.07

0.40

47.85

38.93

FN0193

SVM

28

42.95

0.37

39.20

37.18

FN0132

SVM (LIBRUS)

30

57.37

0.33

25.46

33.55

FN097

Hydrophobicity-probability

5

15.28

0.19

28.57

19.00

FN240

Network centrality

6

14.91

0.06

8.53

10.11

Random Pre

dictor

30

0.10

0.01

0.05

0.06

  1. The third column denotes how many targets in CASP9 are tested in the evaluation of each method. The details for FN0193, FN1032, FN097, and FN240 can be referred to [17].