A fuzzy filter to study the selectivity and sensitivity of a SPME enhanced SAW sensor system characterizing virgin olive oil aroma

An electronic nose based on a solid phase microextration (SPME) unit coupled with a surface acoustic wave (SAW) sensor array has been optimized towards the analysis of virgin olive oil aroma. The parameters of the system, the SPME fibre coating and the kind of carrier gas, have been adjusted to get the optimum response with the minimum relative standard deviation. The selectivity and sensitivity of the sensors were studied by means of a fuzzy filter based on the membership grade functions of the sensor responses for the “non-lampante and “lampante” classes. A multilayer perceptron (MLP) was designed to classify non-lampante and lampante virgin olive oils. Only one sample out of 26 was erroneously classified in the training set, whereas a classification rate of 100% was obtained in a external validation set of 16 samples.