Building honey-based territorial identity for the Formosa Monte through information exploitation using intelligent systems

The territorial valorization of food products is tightly related to quality attributes and is currently the base of typification processes. In the construction of territorial identity based on honey coming from Apis-melifera bee, studies integrating melissopalinological and sensory analysis, and physical-chemical parameters have a significant weight in the definition of the botanical origin and allow the conformation of groups belonging to an area of provenance. The analysis and identification methodologies utilized by classic statistics are based in multivariate techniques, that find their limitations associated to a great number of attributes in a small set of samples. As an alternative, the aim of this article is to define a procedure that allows to determine groups and their descriptive properties, through the exploitation of information using intelligent systems. For the analysis, 47 samples of honey from the north-east region of the oriental district of the Chaqueño Park in Formosa Province, Argentina, were used. A set of nine physical-chemical parameters, 20 sensory descriptors and 128 botanical species (taxa) were also used on these samples. On the first place, affinity groups were identified through the Kohonen algorithm. Afterwards, induction rules were utilized to determine the properties of the groups. Moreover, properties of single-flower (type of dominating pollen) and multi-flower (with more than one kind of pollen intervening) groups of honey were identified through induction rules. The applied techniques resulted in a simplification of the methods employed for honey identification and valorization.