One hundred honey samples of different floral origin (acacia, sunflower, meadow, and forest) collected from nine European countries (Serbia, Albania, Croatia, Montenegro, Romania, Bulgaria, Bosnia and Herzegovina, North Macedonia and Hungary) were analysed for various physicochemical, sensory, antioxidant and antibacterial parameters. The relative antioxidant capacity index and relative antibacterial index were calculated, integrated and expressed as a new property - Power of Honey, intended to be used to predict the health potential of a honey based on its antioxidant and antibacterial activities. Free acidity and colour coordinates L* and a* were chosen for building an artificial neural network model for the prediction of honey health potential. These were chosen based on the obtained correlations between the investigated parameters and in light of the simplicity of the analysis. This model successfully predicted the Power of Honey with a gained coefficient of determination of 0.856. Forest honey samples exhibited the highest Power of Honey. This novel approach should make it possible for honey producers to predict the honey health potential of a particular honey based on a quick and simple analysis.
This journal is © The Royal Society of Chemistry.