Table 3.

Overall accuracy achieved on main dataset, developed using different WEKA methods. Single label was predicted for each substrate and performance evaluated using five-fold cross-validation techniques

Methods	Overall accuracy (percent)
Random Forest	69.47
SMOreg	69.03
Rotation Forest	68.58
Simple logistic	66.37
BayesNet	65.04
REPTree	64.60
RBF network	64.16
Multilayer perceptron	62.39
IB1 (tree)	58.41
NaïveBayes	57.96
KStar (tree)	56.20
Logistic equation	51.77