Partial least square regression (PLSR) is a reference statistical model in chemometrics. In agronomy, it is used to predict components (response variables y) of chemical composition of vegetal materials from spectral near infrared (NIR) data X collected from spectrometers. PLSR reduces the dimension of the spectral data X by defining vectors that are then used as latent variables (LVs) in a multiple linear model. One difficulty is to determine the relevant dimensionality (number of LVs) for the given data. This step can be very time consuming when many datasets have to be processed and/or the datasets are frequently updated. The paper focuses on an alternative, bypassing the determination of the PLSR dimensionality and allowing for automatizing the predictions. The strategy uses ensemble learning methods, such as averaging or stacking the predictions of a set of PLSR models with different dimensionalities. The paper presents various methods of PLSR averaging and stacking and compares their performances to the usual PLSR on six real datasets on different types of forages. The main finding of the study was the overall superiority of the averaging methods compared to the usual PLSR. We therefore believe that such methods can be recommended to analyze NIR data on forages.

This study examined the effects of long-term storage conditions on the chemical composition, pepsin-cellulase dry matter digestibility (PCDMD), and visible (VIS)/near infrared spectra (NIR) of forage. Eighteen samples of different whole-crop maize varieties originally harvested in 1987 were used. After drying, these samples were analyzed in the laboratory for ash, crude protein (CP), structural carbohydrates, total soluble carbohydrates (TSC), starch and PCDMD, and the remaining samples were stored frozen (at −20°C) or at barn temperature (ambient temperatures ranged from −8.5 °C to 27.1 °C). In 2016, the samples were analyzed for ash, CP, structural carbohydrates, TSC, starch and PCDMD. The visible/NIR spectra of both storage methods were obtained. Chemical composition and PCDMD analyses revealed significant differences (p < 0.05) between the storage methods for TSC but not for the other parameters (p > 0.05). After sample harvesting in 1987, the analyses were compared with those in 2016. It was found that the post-harvest TSC and ash content were higher (p < 0.05) and lower (p < 0.05), respectively, during 2016. No significant differences were found for starch and PCDMD. Important differences between the VIS/NIR spectra of both storage methods were obtained in the VIS segment, particularly in the area between 630 and 760 nm. We concluded that storing dry forage samples at ambient temperature for a very long time (29 years) did not change their nutritive value compared to the values obtained before storage.

No abstract