Table 8.

Global confusion matrix obtained using the different classifiers in the case of the Yogev et al. sub-dataset.

		Obtained Classes
		Supervised										Unsueprvised
		K-NN		CART		RF		NB		SVM		K-Means		GMM
		Healthy	PD	Healthy	PD	Healthy	PD	Healthy	PD	Healthy	PD	Healthy	PD	Healthy	PD
True	Healthy	87.50%	12.50%	75.00%	25.00%	81.25%	18.75%	84.38%	15.62%	84.38%	15.62%	71.53%	28.47%	55.63%	44.37%
Classes	PD	14.81%	85.19%	11.11%	88.89%	11.11%	88.89%	31.48%	68.52%	12.96%	87.04%	40.91%	59.09%	29.81%	70.19%