Table 1

Prediction accuracies for decision trees (C5.0), SVMs, and Random Forests.

Classifier accuracy

Single Seq (5000 samples)

HSSP Profiles (5000 samples)

Property profiles (5000 samples)

Property profiles (20000 samples)

C5.0

-

0.2320

-

-

SVM

0.2955

0.3038

0.3426

0.3615

SVM (prob. estim.)

-

-

0.2900

-

Random Forest

0.3024

0.3001

0.3102

0.3409

Sander et al. BMC Bioinformatics 2006 7:14   doi:10.1186/1471-2105-7-14