Table 3.

Performance metrics for the individual classifiers (TD Group: ID1–ID13, ASD Group: ID14–ID18).

ID	Interaction length (s)	No. of datapoints (frames)	Train		Validation		Test
			Accuracy	Loss	Accuracy	Loss	Accuracy
1	315	9,444	0.8101	0.5028	0.7790	0.6681	0.7946
2	519	15,357	0.6499	0.7278	0.6398	0.7797	0.6393
3	540	16,412	0.6703	0.8723	0.6407	1.0095	0.6526
4	658	10,933	0.8302	0.4189	0.8131	0.4923	0.8240
5	797	22,996	0.9255	0.1903	0.9198	0.2484	0.9159
6	696	9,300	0.9200	0.2850	0.8925	0.3856	0.9124
7	316	9,388	0.7821	0.5423	0.7417	0.7946	0.7338
8	457	13,725	0.7561	0.6065	0.7418	0.6796	0.7483
9	574	10,463	0.6671	0.8486	0.6535	0.9333	0.6364
10	780	16,627	0.9104	0.2253	0.8831	0.3907	0.8698
11	726	12,726	0.8390	0.3843	0.8303	0.4039	0.8283
12	685	9,723	0.8118	0.5162	0.7715	0.6980	0.7720
13	540	12,879	0.8084	0.4296	0.7812	0.5858	0.7702
14	517	15,502	0.8163	0.4417	0.7952	0.5621	0.7907
15	578	14,624	0.9204	0.2276	0.8923	0.3390	0.9108
16	679	15,950	0.6810	0.7582	0.6501	0.9095	0.6398
17	610	16,401	0.8306	0.3946	0.8232	0.4923	0.8366
18	1058	30,508	0.7822	0.5467	0.7759	0.6323	0.7812