3.4. Prediction of different characteristics of beef muscle using partial least
square regression
The results of PLS regression with cross validation of the three
muscles to gather (ST + RA + IS) showed poor prediction for collagen
(R2C= 0.71, R2V= 0.25) from the emission spectra collected in
the range of 305–400 nm (Table 2). However, good prediction was
obtained when the regression model is applied on the individual
muscle.
Good results for prediction of energy required for rupture and dry
matter content (high R2C) were also observed from the emission spectra
collected in the range of 410–700 nm and 340–540 nmrespectively.
However, high R2V was only observed for spectral range 340–540 nm
and drymatter content of ST, RA and IS muscles (Table 2). This spectral
range might correspond to the vitamin A present in the meat sample
(Skjervold et al., 2003). As vitamin A is located mainly in fat (Dufour &
Riaublanc, 1997), so it is speculated that there may be correlations
between vitamin A fluorescence spectra and dry matter contents of
meat.