Real-Time Detection of Honey Adulteration Using a Portable E-Nose System

Honey adulteration poses a significant threat to food quality and consumer health, necessitating the development of rapid and cost-effective detection methods. This study presents a portable electronic nose (e-nose) system equipped with a cross-selective array of metal oxide semiconductor (MOS) sensors for detecting adulterated honey based on volatile organic compound (VOC) patterns. The e-nose responses for 120 honey samples were analyzed using Principal Component Analysis (PCA) for feature reduction, followed by classification with supervised machine learning models. The Support vector machine (SVM) classification model achieved better results with a classification accuracy of 91.9%, with precision of 92.2%, recall of 92.5%, and an f1-score of 92.3%, effectively distinguishing pure honey from jaggery-adulterated samples. The results demonstrate that identifying individual VOCs is not essential; instead, analyzing the collective VOC signature using cross-response sensors and chemometric modeling provides a powerful, low-cost, and non-invasive alternative to traditional analytical methods like GC-MS. This work establishes the proposed e-nose system as a promising tool for real-time honey adulteration detection in food quality control applications.