Table 4. Classifiers performance of PSD-based feature extraction with oversampling using SAM and HAM-A labeling.

	HAM-A label			SAM label
	Accuracy	Recall	Precision	Accuracy	Recall	Precision
KNN	70.09%	70.09%	72.59%	68.04%	68.04%	69.26%
LDA	70.74%	70.04%	72.25%	62.27%	62.27%	63.03%
SVM	67.50%	67.50%	68.63%	59.15%	59.15%	59.89%
RF	84.07%	84.07%	84.15%	79.40%	79.40%	79.50%
AdaBoost bagging	76.11%	76.11%	76.29%	69.32%	69.32%	69.41%
Gradient Bagging	76.20%	76.20%	76.37%	69.14%	69.14%	69.25%