Table 3 Model performance and statistical comparisons of video featuring paired macaques (Experiment 2). Bolded p-values indicate significant at \(p<0.05\).

Model	Dataset	Precision	Recall	F1
Model performance
MacqD - Macaque Single	‘Same’	0.79	0.66	0.71
MacqD - Macaque Curriculum		0.94	0.87	0.90
MacqD - Macaque Combine		0.97	0.81	0.87
MacqD - Macaque Single	‘Different’	0.72	0.54	0.60
MacqD - Macaque Curriculum		0.89	0.72	0.81
MacqD - Macaque Combine		0.95	0.70	0.81
BE		0.45	0.40	0.42

Comparison	Dataset	N	DOF	Metric of comparison	\(\chi ^{2}\)	p-value
Friedman test statistics
MacqD - Macaque Single versus MacqD - Macaque Curriculum versus MacqD - Macaque Combine	‘Same’	10	2	Precision	8.16	0.017
				Recall	6.53	0.038
				F1	7.59	0.023
MacqD - Macaque Single versus MacqD - Macaque Curriculum versus MacqD - Macaque Combine versus BE	‘Different’		3	Precision	25.56	0.000
				Recall	10.16	0.017
				F1	11.42	0.010

Comparison	Dataset	N	DOF	Metric of comparison	\(T^{+}\)	Corrected p-value
Wilcoxon test statistics
MacqD - Macaque Single versus MacqD - Macaque Curriculum	‘Same’	10	1	Precision	6.5	0.041
				Recall	5	0.059
				F1	2	0.038
MacqD - Macaque Single versus MacqD - Macaque Combine				Precision	0.0	0.035
				Recall	14.5	0.193
				F1	3.5	0.038
MacqD - Macaque Curriculum versus MacqD - Macaque Combine				Precision	20.5	0.812
				Recall	6	0.139
				F1	7.5	0.141
MacqD - Macaque Single versus MacqD - Macaque Curriculum	‘Different’			Precision	4.5	0.016
				Recall	12	0.157
				F1	8.5	0.073
MacqD - Macaque Single versus MacqD - Macaque Combine				Precision	0	0.004
				Recall	22	0.625
				F1	18.5	0.441
MacqD - Macaque Curriculum versus MacqD - Macaque Combine				Precision	11	0.105
				Recall	7	0.157
				F1	16	0.441
MacqD - Macaque Single versus BE				Precision	1	0.006
				Recall	9	0.157
				F1	3	0.041
MacqD - Macaque Curriculum versus BE				Precision	0	0.002
				Recall	7	0.157
				F1	4	0.041
MacqD - Macaque Combine versus BE				Precision	0	0.002
				Recall	11	0.157
				F1	6	0.055

Search