偏最小二乘回归
高光谱成像
掺假者
主成分分析
肌内脂肪
食品科学
化学
化学计量学
数学
色谱法
人工智能
计算机科学
统计
作者
J.P. Cruz‐Tirado,Maria do Carmo Vieira,Oscar Oswaldo Vásquez Correa,Daphne Ramos D.,José M. Angulo-Tisoc,Douglas Fernandes Barbin,Raúl Siché
标识
DOI:10.1016/j.jfca.2023.105901
摘要
Alpaca meat has high protein content; good tenderness and low intramuscular fat content, being more expensive than meat from other animals (e.g., beef). Hence, alpaca meat may face adulteration, which demands chemical analytical methods to be identified. In this study, chemical-free methods, e.g., portable NIR spectrometer and NIR-HSI were employed to detect adulteration of alpaca meat with pork, chicken, and beef (0 – 50% w/w). Spectral analysis revealed significant differences in the spectra of pure alpaca meat samples using NIR-HSI. Principal Component Analysis (PCA) grouped samples into pure and non-pure alpaca meat using both devices as sources of spectra. Next, we developed and validated one-class Data Driven Soft Independent Class Analogy (DD-SIMCA) models to authenticate pure alpaca meat. DD-SIMCA models using spectra acquired by both devices achieved 100% sensitivity and 100% specificity for external set of samples. Besides, Partial Least Squares Regression (PLSR) based on NIR-HSI outperformed the portable NIR spectrometer to predict the concentration of adulterant in alpaca meat. In conclusion, both devices supported by chemometric approaches can be implemented as screening methods to detect adulteration in alpaca meat.
科研通智能强力驱动
Strongly Powered by AbleSci AI