Document Type : Research Article
Authors
1
department of Food Science & Nutrition, medical school, Gonabad university of medical sciences, Gonabad, Iran
2
Halal Research Center of IRI, Iran Food and Drug Administration, Ministry of Health and Medical Education, Tehran, Iran
3
Department of General Courses, Faculty of Medicine, Gonabad University of Medical Sciences, Gonabad, Iran
4
food safety and hygiene department, school of health, university of medical science, Gonabad, Iran.
Abstract
Rapid and low-waste assessment of grape quality remains a challenge for conventional analytical techniques; Raman spectroscopy combined with chemometric modeling provides a promising alternative for predicting the quality attributes of Asgari grapes. Fifty-five grape samples were collected from vineyards across Gonabad, Iran, and key physicochemical parameters—including pH, titratable acidity (TA), flavonoid content, total soluble solids (TSS), anthocyanin content, and the TSS/TA ratio—were measured. Dimensionality reduction was carried out using principal component analysis (PCA) and partial least squares (PLS), followed by the development of predictive models using multiple linear regression (MLR) and support vector machines (SVM). The dataset was randomly divided into training (80%) and testing (20%) subsets. Model performance was evaluated using the coefficient of determination (R²) and the root mean square error of calibration (RMSEC) and prediction (RMSEP). Among the evaluated models, the PCA–MLR hybrid demonstrated the best predictive performance, achieving R² values of 0.75 for anthocyanins, 0.74 for flavonoids, and 0.65 for TSS, with corresponding RMSEC values of 0.56, 0.67, and 1.51, respectively. These results confirm that Raman spectroscopy is a viable and efficient tool for grape quality assessment and represents a practical alternative to conventional chemical analytical methods.
Keywords
Main Subjects
)