Skip to main content
. 2013 Sep 23;29(24):3135–3142. doi: 10.1093/bioinformatics/btt554

Table 2.

Results for the ensemble prediction of IVB effectors with 10-fold cross-validation tests

Classifiers Modela Votingb Sn Sp Acc MCC
  • 1. AAC

  • 2. DPC

  • 3. PSSM

  • 4. PSSM_AC

1 0.707 0.933 0.884 0.651
2 0.658 0.928 0.870 0.603
3 0.903 0.971 0.956 0.871
4 0.839 0.959 0.933 0.800
{1,2,3} 2-in-3 0.832 0.990 0.956 0.867
{1,2,4} 2-in-3 0.816 0.984 0.948 0.842
{1,3,4} 2-in-3 0.916 0.996 0.979 0.936
{2,3,4} 2-in-3 0.894 0.994 0.972 0.917
{1,2,3,4} 3-in-4 0.774 0.977 0.933 0.796

aEach term in this column represents the combination form of different classifiers (e.g. {1,2,3} means the ensemble of number 1, 2 and 3 classifiers in the first column).

bThe 2-in-3 voting scheme was adopted for voting on three classifiers and the 3-in-4 voting scheme was adopted for voting on four classifiers. The highest value of each measure is shown in bold.