Categorising cheetah behaviour using tri-axial accelerometer data loggers: a comparison of model resolution and data logger performance

Table 3 Mean cross-validation and training data metrics

Model	Device	Training				Cross-validation
Model	Device	Accuracy	RMSE	MSE	r²	Accuracy	RMSE	MSE	r²
Fine-scale	CEFAS	0.825	0.424	0.179	0.987	0.825 (0.005)	0.431 (0.004)	0.186 (0.003)	0.986 (< 0.001)
	GCDC	0.846	0.416	0.173	0.987	0.843 (0.003)	0.424 (0.001)	0.180 (0.001)	0.987 (< 0.001)
Medium-scale	CEFAS	0.838	0.398	0.158	0.982	0.837 (0.003)	0.404 (0.003)	0.163 (0.003)	0.981 (< 0.001)
	GCDC	0.852	0.394	0.155	0.982	0.849 (0.004)	0.401 (0.001)	0.161 (0.001)	0.981 (< 0.001)
Coarse-scale	CEFAS	0.877	0.327	0.107	0.885	0.876 (0.003)	0.329 (0.003)	0.109 (0.002)	0.883 (0.002)
	GCDC	0.894	0.312	0.097	0.894	0.890 (0.002)	0.319 (0.001)	0.102 (0.001)	0.889 (0.002)

ISSN: 2051-3933