Vol. 10, Issue 4 (2021)

Author(s):
Mrinmoy Roy, N Hariharan, Manoj D, Ishita Auddy and S Shanmugasundaram

Abstract:
The cow ghee is derived from milk and contains high-quality nutrients. The adulteration of cow ghee is very common among traders to earn more profit. The traditional methods of adulteration detection require sample pre-processing and are also time-consuming and costly processes. Therefore, machine olfaction instruments are very common now a day to evaluate the quality of foodstuffs in a fast and precise way. The potential of metal oxide semiconductor (MOS) gas sensor based electronic nose (e-nose) system was developed to detect adulteration in ghee with hydrogenated fat (vanaspati). The data obtained from the e-nose system were analyzed for pattern recognition and classification using multivariate chemometric analysis such as principal component analysis (PCA) and discriminant function analysis (DFA) method for adulteration detection in ghee. The PCA explained 98.10% variance while, DFA explained 99.10% variance in the e-nose dataset. The accuracy in training data and its cross-validation was found to be 98.18% and 97.27%, respectively. The success rate of adulteration identification for test samples using DFA model was 90.90% based on received e-nose signals. The results of PCA and DFA suggest that the developed e-nose system successfully identified pure and adulterated ghee samples based on the e-nose data.