Table 3

pairwise comparison of predictions made by FAM and SVM for 369 GO functions in 4 cross-validation experiments. For each of the 369 GO functions, the performance of FAM and SVM are compared. "Win," "Tie" and "Lose" refer to the number of GO functions for which FAM's predictions are better than, tied with, or worse than that of SVM, respectively. The last row, denoted by "ACV," compares FAM and SVM using performance metrics which are averaged over the 4 cross-validation folds. As expected, this gives rise to fewer ties.

CV Fold
Sensitivity10%
Specificity10%
AUC

Win
Tie
Lose
Win
Tie
Lose
Win
Tie
Lose

1
106
168
95
107
167
95
193
2
174
2
87
181
101
88
183
98
177
1
191
3
104
171
94
109
166
94
186
3
180
4
90
193
86
92
192
85
199
6
164
ACV
184
36
146
194
34
141
201
0
168

Kustra et al. BMC Bioinformatics 2006 7:216   doi:10.1186/1471-2105-7-216