Chemical-instrumental-sensory traits and data mining for classifying dry-cured Iberian shoulders from pigs with different diets

Abstract

Iberian pigs are an autochthonous porcine breed exclusively from the south western of Iberian Peninsula. In this study, the main objective was to classify dry-cured Iberian shoulders from pigs with different diets. Thus, morphology, physico-chemical and sensory parameters, fatty acid profile and volatile compounds were determined. From this data, two datasets were created, for training and validation purpose. Results on this study, firstly demonstrate the capability of data mining techniques to classify shoulder as function of their different diets by using different chemical-instrumental-sensory parameters. Different classification models were tested in the training datasets. After that, all classification models were performed in the validation datasets and the model of J48 decision tree and fatty acid profile reached the best results (Sensitivity and Specificity > 0.750). From this classification model, a software application was developed for determining the diet of the Iberian pigs. This application could be used for the meat industries and inspection agencies.