TABLE III.

AUD Dataset. Top: Comparison across methods on NC vs. AUD classification; Bottom: Balanced Classification Accuracy of the proposed method dependent on the number of visits

Method	BACC ± std	SEN	SPE	AUC
CNN	67.6 ± 3.54	62.9	72.2	70.9
CNN+AP	68.7 ± 3.68	66.7	70.7	71.5
CNN+RNN	69.0 ± 3.35	67.1	70.9	71.6
CNN+RNN+LP	69.5 ± 3.32	67.6	71.4	71.8
CNN+RNN+LP+CL	69.3 ± 3.31	67.9	70.8	71.8
CNN+biRNN [7]	68.9 ± 3.36	67.8	70.0	71.7
CNN+tRNN [27]	69.2 ± 3.42	68.6	69.8	71.8

	Visits	NC/AUD	BACC ± std
CNN+RNN+LP+CL	1+ visit	244/293	65.0 ± 3.34
	2+ visits	124/150	70.6 ± 3.32
	3+ visits	63/74	77.1 ± 3.28
	4+ visits	30/29	79.5 ± 3.29