‘Artificial Sensory Analysis’ for sensory classification of Prosecco sparkling wines
TIRANNO M., FRANCESCHI D., VINCENZI S., BOATTO V., BRAVI M.; C. I. R. V. E. Centro Interdipartimentale per la Ricerca in Viticoltura ed Enologia, Univ
Market, technology and scientific motivations support the development of devices for carrying out sensory evaluations by instrumental techniques.
An experimentation aimed at develop a classification and nonconformity detection tool for ‘Prosecco’ (DOC) and ‘Prosecco Superiore’ (DOCG) sparkling wines is presented here.
All the anonymised samples which are classed as conformant/nonconformant and belonging to the strongly/weakly aromatic by routine panel testing are also passed through an experimental rig based on a commercial electronic nose according to a in-house defined protocol.
The input data to the classification algorithm include the filtered e-nose output and some gas phase-relevant chemical/physical data analysis results on the same wine, construction of the loading plot of the sample data.
The preliminary results after three months of routine analysis show that a probabilistic mechanistic aromatic classification scheme is possible. Disciplined products could be classified by subjecting only a fraction of the total samples to the panel analysis, depending upon the results of the preliminary classification performed by the enose. All samples classed by the algorithm as ‘likely non conformant’ would be panel-tested while only a fraction of the ‘likely conformant’ samples would be panel-tested, thereby permitting an increased panel throughput while maintaining the ability of the instrumental system to perform the classification according to the desired accuracy.
Poster presented at Enoforum 2013, 7-9 May, Arezzo (Italy)
Published on 05/24/2015
- (1,444 Mb)
Open and read the article (pdf, 428 Kb)
- (2,722 Mb)
Open and read the poster (pdf, 2.0 Mb)