Premium
Characterization of olive oil classes using a Chemsensor and pattern recognition techniques
Author(s) -
Peña F.,
Cárdenas S.,
Gallego M.,
Valcárcel M.
Publication year - 2002
Publication title -
journal of the american oil chemists' society
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.512
H-Index - 117
eISSN - 1558-9331
pISSN - 0003-021X
DOI - 10.1007/s11746-002-0611-6
Subject(s) - olive oil , principal component analysis , pattern recognition (psychology) , artificial intelligence , computer science , quality (philosophy) , quality assurance , mathematics , food science , chemistry , engineering , philosophy , operations management , external quality assessment , epistemology
Classification is an important component of food quality assurance, as methods to guarantee authenticity of food products are widely demanded by food producers, processors, consumers, and regulatory bodies. The objective of this work was to develop a rapid classification method in order to discriminate virgin olive oil, olive oil, and „orujo” olive oil, the prices of which differ dramatically in the market on a ccount of the high quality level of the former. For these purposes, new ChemSensor equipment that combines a headspace autosampler with a mass‐selective detector and Pirouette data evaluation software was used. To take into account the large number of samples analyzed (50 samples repeated 10 times), as well as the wide interval of m/z ratios scanned (41–170), chemometric approaches were necessary. Cluster analysis, principal component analysis, K‐nearest neighbors, and soft independent modeling of class analogy (SIMCA) were applied to model the different oil classes. The results indicated good classification and prediction abilities, with SIMCA affording the best results ( viz . 97% specificity).