Quantitative and qualitative evaluation of maltodextrin products in the industry using near-infrared spectroscopy

Abstract Maltodextrin is a crucial ingredient in food and pharmaceutical sectors. Traditional quality assessment methods for maltodextrin are destructive and time consuming. This study aimed to employ near-infrared (NIR) spectroscopy and chemometrics to differentiate maltodextrin variants and measure their quality parameters efficiently. NIR spectra were recorded in the range of 12 000–4000 cm−1 using the transflectance mode. The classification model effectively distinguished maltodextrin types based on their dextrose equivalent (DE) values using techniques such as partial least squares-discriminant analysis and supervised self-organising map (SSOM). Moreover, quality parameters including moisture content, DE value, maltose, maltotriose, pH, and SO2 were quantitatively assessed using partial least squares regression (PLSR) and SSOM models. Particularly, PLSR provided better results, with residual predictive deviation values exceeding 2.5 for moisture content, DE values, maltose, and maltotriose. These models can be applied for use in both laboratory settings and industrial monitoring.