Discrimination of Corsican honey by FT-Raman spectroscopy and chemometrics

Abstract

Honey is a complex and challenging product to analyze due mainly to its composition consisting on various botanical sources. The discrimination of the origin of honey is of prime importance in order to reinforce the consumer trust in this typical food product. But this is not an easy task as usually no single chemical or physical parameter is sufficient. The aim of our paper is to investigate whether FT-Raman spectroscopy as spectroscopic fingerprint technique combined with some chemometric tools can be used as a rapid and reliable method for the discrimination of honey according to their source. In addition to that, different chemometric models are constructed in order to discriminate between Corsican honeys and honey coming from other regions in France, Italy, Austria, Germany and Ireland based on their FT-Raman spectra. These regions show a large variation in their plants. The developed models include the use of exploratory techniques as the Fisher criterion for wavenumber selection and supervised methods as Partial Least Squares-Discriminant Analysis (PLS-DA) or Support Vector Machines (SVM). All these models showed a correct classification ratio between 85% and 90% of average showing that Raman spectroscopy combined to chemometric treatments is a promising way for rapid and non-expensive discrimination of honey according to their origin.

Keywords

• Discrimination
• FT-Raman
• typical food product
• honey
• chemometrics
• SVM
• PLS-DA