Chemometric characterization of the claretes and rose wines of the certified denomination of origin Rioja using CieLab parameters

Multivariate chemometric techniques such as SIMCA (Soft Independent Modelli ng Class Analogy) and GINN (Genetic Inside Neural Network) were used to con struct sensitive and specific models for rose and 'claretes' wines of the C ertified Denomination of Origin Rioja, on the basis of data obtained from a n easy espectrophotometric analysis (Cie-Lab parameters). GINN used as mode lling method provided an overall percentage of success over 99.26% in all c lasses. The model built with GINN has high sensitivity and specificity: 100 % and 96.43% for 'claretes' and rose wines. While the SIMCA models have a s lightly lower percentage of sensitivity, the model for rose wines does not admit any wine from the evaluation set formed with claretes and blended win es.