Table 2.

Temperature prediction (preanalysis). Coefficients of determination ($R^2$) illustrating the predictive power of temperature prediction based on classical least squares (CLS), inverse least squares (ILS), and principal component regression (PCR) (one to three PCs), for the animal and the human datasets. Values represent results averaged across the animal and the human dataset.

	Animal dataset				Human dataset
	740 nm	840 nm	970 nm	720 to 880 nm	840 nm
	CLS				CLS
$R^2$	0.231	0.299	0.034	0.383	0.670
	ILS				ILS
$R^2$	0.538	0.609	0.077	0.684	0.697
	PCR				PCR
$R^2$ (one PC)	0.593	0.688	0.112	0.690	0.788
$R^2$ (two PCs)	0.593	0.689	0.111	0.706	0.789
$R^2$ (three PCs)	0.590	0.684	0.111	0.706	0.790