Identification of geographical and botanical origin of Mediterranean honeys using UV-vis spectroscopy and multivariate statistical analysis

The present study aimed to differentiate Mediterranean honey samples with regards to their botanical and geographical origin, using UV-vis spectroscopy, combined with multivariate statistical analysis and chemometrics. A total of 151 commercial honey samples were collected from different countries (Greece, Malta, Spain, Tunisia, and Turkey), which were characterized as thyme, pine, or polyfloral. Discrimination of geographical and botanical origin was achieved successfully with Principal Component Analysis (PCA) together with Random Forest (RF), Partial Least Squares - Discriminant Analysis (PLS - DA), and Data Driven - Soft Independent Modelling of Class Analogies (DD-SIMCA) using the spectral range of 220-550 nm. DD-SIMCA achieved very high accuracy, sensitivity, and specificity on the discrimination of both the geographical and botanical origin of honey samples. The results support the good predictive capability of UV-vis spectroscopy combined with chemometrics for the determination of honey origin. This methodology could be utilized for the rapid and non-destructive assessment of Mediterranean honey origin.