We found a match
Your institution may have access to this item. Find your institution then sign in to continue.
- Title
Miniaturized hand-held near-infrared spectroscopy and machine learning for precision monitoring of solid fat content.
- Authors
Kucha, Christopher; Olaniyi, Ebenezer O.; Ngadi, Michael
- Abstract
The functionality of fats in food is determined by the total amount of fat crystals, measured as percent solid fat content (SFC). The SFC is a physical parameter conventionally evaluated by either nuclear magnetic resonance or differential scanning calorimetry (DSC), which is time-consuming. This study investigated the use of near-infrared (NIR) spectroscopy for the prediction of SFC in fats. Fat and oil blends consisting of lard, tallow, peanut oil, palm oil, and canola oil were analyzed for values of SFC (%) at 0, 10, 20, and 30 °C by using the DSC. The spectra were obtained using a NIR spectroscopy (1159–2169 nm). A partial least square regression (PLSR) algorithm was used to build correlation models based on full wavelengths. Successive projection algorithm (SPA), and regression coefficients (RCs) were used to select effective wavelengths to build new PLSR models. An independent sample set was used to validate the models. The models were evaluated in terms of coefficient of determination of prediction R p 2 and root mean square error of prediction (RMSEP). The optimal model based on the full spectra (125 variables) showed accurate results ( R p 2 = 0.96, RMSEP = 2.16 %) in the prediction dataset at 20 °C, while the new models equally yielded good correlation for the SPA (14 variables) ( R p 2 = 0.92, RMSEP = 2.63%) and RC (10 variables) ( R p 2 = 0.93, RMSEP = 2.8%) based PLSR models at 10 °C and 20 °C respectively. The promising result in this study shows the potential of NIR spectroscopy to be a good alternative to the time-consuming methods for monitoring SFC in fat. The selected wavelengths could be used to develop a multispectral spectrometer for an industrial application.
- Subjects
NEAR infrared spectroscopy; MACHINE learning; STANDARD deviations; NUCLEAR magnetic resonance; FAT; PEANUT oil
- Publication
Journal of Food Measurement & Characterization, 2024, Vol 18, Issue 6, p4417
- ISSN
2193-4126
- Publication type
Article
- DOI
10.1007/s11694-024-02504-y