Table 5. First scenario classification accuracies under different feature subsets and classifiers.

Classifier	Accuracy using All features (51)	Accuracy with feature selection	No. of selected features	Accuracy using features with significance statistical P-value	No. of selected features
Linear SVM	93.0 ± 0.16 %	97.0 ± 0.13 %	9	97.0 ± 0.13 %	11
Random forest	90.0 ± 0.27 %	97.0 ± 0.13 %	11	97.0 ± 0.13 %	15
Adaboost	85.0 ± 0.40 %	93.0 ± 0.16 %	11	90.0 ± 0.24 %	21
MLP	93.0 ± 0.16 %	97.0 ± 0.13 %	20	97.0 ± 0.13 %	22
Linear via SGD	90.0 ± 0.16 %	97.0 ± 0.13 %	14	97.0 ± 0.13 %	21
Mean	90.2 %	96.2 %	13	95.6 %	18