Application of numerical classification for recognition of floral and sugar-feeding bee honey

The objective of the present study is to use discriminant analysis to distinguish three types of floral honey from each other and from sugar-feeding honey by parameters that can be determined easily in routine honey control. The characterization of three types of floral (citrus, clover, cotton) and sugar-feeding honey was carried out based on their quality parameters [pH, glucose, fructose, sucrose, maltose content, total soluble solids (TSS), total acidity, free acids, lactones, electrical conductivity]. All the data were statistically tested using analysis of variance and discriminant analysis, which was used in this study as pattern recognition, as a classification tool and to identify the most important parameters in the classification. The results showed that citrus, clover, and cotton honey could not be assigned by 100% into their actual groups even when all parameters were used simultaneously in the analysis. Two samples of the floral honey were assigned as sugar-feeding honey. Among different parameters used, TSS, total acidity, free acidity, fructose, glucose, fructose content, EC accounted for the most distinguishing parameters between the different honey types. The results were discussed in the light of some beekeeping managements before and after harvesting of honey, and the effectiveness of the chosen parameters.