Monitoring the freshness of Moroccan sardines with a neural-network based electronic nose

An electronic nose was developed and used as a rapid technique to classify the freshness of sardine samples according to the number of days spent under cold storage ( 4 +/- 1 degrees C, in air). The volatile compounds present in the headspace of weighted sardine samples were introduced into a sensor chamber and the response signals of the sensors were recorded as a function of time. Commercially available gas sensors based on metal oxide semiconductors were used and both static and dynamic features from the sensor conductance response were input to the pattern recognition engine. Data analysis was performed by three different pattern recognition methods such as probabilistic neural networks (PNN), fuzzy ARTMAP neural networks (FANN) and support vector machines (SVM). The objective of this study was to find, among these three pattern recognition methods, the most suitable one for accurately identifying the days of cold storage undergone by sardine samples. The results show that the electronic nose can monitor the freshness of sardine samples stored at 4 degrees C, and that the best classification and prediction are obtained with SVM neural network. The SVM approach shows improved classification performances, reducing the amount of misclassified samples down to 3.75%.