Recognition of beef aging time using a miniaturized near-infrared spectrometer in tandem with support vector machine

Abstract

Consumers increasingly demand sustainable production practices and high-quality standards. Near-infrared (NIR) spectroscopy presents a non-invasive and efficient tool for addressing these concerns. This study aimed to evaluate vacuum-aged beef across different aging periods (3, 10, 17, and 24-days) using NIR spectroscopy and chemometric methods, focusing on a classification model based on SVM. NIR spectra were collected from 356 samples, and PCA was performed. The first and third principal components explained 83.05 % of the variance, showing grouping tendencies for samples aged 3 and 10 days versus those aged 17 and 24 days. Chemical groups related to aging, such as proteins, water, lipids, acids, and alcohols, drove spectral differentiation. The spectral region between 1228 and 1337 nm was identified as the most relevant for model development, achieving an accuracy of 96.60 %. This study demonstrates that portable NIR spectrometer, in combination with SVM classification, offer a fast, cost-effective, and user-friendly method for classifying aged beef.