Predicting the chemical composition and in vitro digestibility properties of fermented product produced from maize harvest residue utilising near infrared reflectance spectroscopy

Near infrared (NIR) spectroscopy enables simultaneous rapid assessment of multiple parameters. Partial least squares regression of NIR spectra was applied to identify multiple fermented maize product quality components (the crude protein (CP), ether extract (EE), crude fibre (CF), neutral acid detergent fibre (NDF), acid detergent fibre (ADF), acid detergent lignin (ADL), neutral detergent insoluble protein (NDIP), acid detergent insoluble protein (ADIP), hemicellulose, energy parameters (metabolic energy (ME), net energy lactation (NEL), digestibility parameters (ammonia nitrogen (NH 3 /N), in vitro dry matter digestibility (IVDMD), in vitro neutral detergent fibre digestibility (IVNDFD), organic matter digestibility (OMD), and gas production parameters. This study aimed to develop a unique NIR spectroscopy calibration method for fermented maize harvest residue (FMHR) utilising 324 samples with various FMHR additions. The coefficients of determination (R 2 ) in the calibration set showed the best results for all quality, energy, and digestibility parameters, except for EE, CF, and NDIP. Gas production of the insoluble fraction, gas production at 24 h, and total gas production (96 h) indicated the best prediction results among the biological parameters related to gas production. The equations for IVDMD (R 2 = 0.99), CP (R 2 = 0.97), and NH 3 /N (R 2 = 0.97) showed high accuracies. The equations for IVNDFD (R 2 = 0.95) and OMD (R 2 = 0.93) showed relatively high accuracies. It was demonstrated that it is a feasible technique for monitoring feed quality and digestibility parameters in maize harvest residues by the developed model in NIR spectrophotometer, thereby aiding managing and optimizing diets. In the future, widening the data range of the initial dataset could further improve the accuracy of these models.