Table 2. Evaluation parameters from the prediction of bioactivity class of an independent NAA test dataset by the four classifier models used in this study.

Classifier Models	Accuracy (%)	Kappa Statistics	Area Under Curve (ROC)
Random Forest (RF)	82.81	0.65	0.91
Voted Perceptron (VP)	71.48	0.42	0.72
Sequence Minimization Optimization (SMO) of Support Vector Machine	85.94	0.72	0.86
Naïve Bayesian (NB)	73.05	0.45	0.74
Fused Model (RF and SMO)	82.03	0.68	0.92

Comments: RF: Random forest of 10 trees, each constructed while considering 11 random features. Out of bag error: 0.1797. SMO: The polynomial kernel. Fused Model (RF and SMO): The predictions from RF and SMO were combined (mean) using the “Prediction Fusion” node in KNIME [38].