Table II.

Classification performance of the models developed with and without the ML in the case of the UDIAT and OASBUD datasets. The standard deviations of the parameters were calculated using bootstrap.

Dataset	Method	AUC	Accuracy	Sensitivity	Specificity
UDIAT	MP features	0.858±0.029	0.853±0.024	0.796±0.043	0.880±0.027
	FC features	0.849±0.031	0.822±0.037	0.759±0.043	0.853±0.053
	MP features, ML	0.873±0.027	0.840±0.023	0.833±0.037	0.844±0.027
	FC features, ML	0.893±0.030	0.840±0.024	0.851±0.042	0.834±0.030
OASBUD	MP features	0.819±0.030	0.760±0.029	0.692±0.057	0.833±0.057
	FC features	0.791±0.035	0.750±0.031	0.750±0.044	0.750±0.050
	MP features, ML	0.831±0.031	0.760±0.031	0.762±0.059	0.750±0.061
	FC features, ML	0.881±0.023	0.830±0.026	0.807±0.039	0.854±0.036