NIR spectroscopy and PLS regression with waveband selection for estimating chlorogenic acid content of perennial ryegrass
2019
Eguchi, K. (Central Region Agricultural Research Center, NARO (Japan)) | Kiyoshi, T. | Kimura, T. | Kawamura, K.
Perennial ryegrass (Lolium perenne L.) is valuable for use in cool seasonal lawns and is widely used as fodder for grazing and forage crops. It is rich in chlorogenic acid as a functional ingredient. In the future, it is likely that breeding with a focus on this acid may be established. Hence, the development of a simple and rapid method for the estimation of chlorogenic acid content is required. In this study, a calibration curve of the chlorogenic acid content of perennial ryegrass was obtained using near-infrared spectroscopy. Subsequent analysis of the data was based on the partial least squares method (PLSR) with full-spectra information to contribute to the estimation of the target components. The iterative stepwise elimination PLSR (ISE-PLSR) method and genetic algorithm PLSR (GA-PLSR), as a wavelength selective PLSR method for eliminating non-existent wavelength information or selecting useful wavelength information, were also utilized. The estimation accuracies of the approaches were compared using the determination coefficient (R**2) and the residual prediction deviation (RPD). The ISE-PLSR method based on the second derivative absorbance (SDA) spectrum resulted in an R**2 of 0.872 and an RPD of 2.728. These values were more accurate compared to the other methods (R**2 = 0.803 - 0.847, RPD = 2.309 - 2.533). Because RPD> 2.43, it can be concluded that the ISE-PLSR model using the SDA spectrum exhibited high accuracy.
Показать больше [+] Меньше [-]