Authentication of Romanian honeys based on physicochemical properties, texture and chemometric

The purpose of this study was to investigate the physico-chemical properties (free acidity, pH, a_w, ash content, moisture content, color (L*, a*, b*, hue-angle, chroma and yellow index), fructose, glucose and sucrose content) and textural parameters (viscosity, hardness, adhesion, springiness, cohesiveness, chewiness and gumminess) of 50 samples of honey of different botanical origin (acacia, polyfloral, honeydew, sunflower and tilia). In order to achieve the authentication of the honey samples analyzed, their data have been subjected to linear discriminant analysis (LDA) and principal component analysis (PCA).The PCA and LDA have proved the possibility of honey authentication using the physico-chemical and textural properties. LDA classified correctly 92.0% of the honeys based on their botanical origin, using the cross validation. In the LDA projection, the textural parameters (chewiness, hardness, cohesiveness, springiness) dominated the two functions.