Figure 2.
Jackknife cross-validation and independent testing accuracy of the random forest classifier with the number of features varied: (A) 2-gap dipeptide composition (2-gapDC) features (B) 59 selected 2-gapDC features + 60 split amino acid composition (SAAC) features, and (C) 55 selected 2-gapDC features + 60 SAAC features.