Abstract:

This study attempted to use an electronic nose (PEN3) to discriminate the strains of lactic acid bacteria in goat
yogurt samples. The volatile components emanating from goat yogurt samples were gathered by the electronic nose. Based
on the data obtained, a method for discriminating the strains of lactic acid bacteria in goat yogurt was established through
principal component analysis (PCA), Fisher linear discriminant analysis (FLDA) and BP neural network. The results showed
that although both PCA and FLDA could discriminate different species of lactic acid bacteria, FLDA was more effective
than PCA. The correct prediction rates of FLDA and BP neural network were 100.0% and 98.4%, respectively. These results
will be helpful for the application of electronic nose to discriminate the strains of lactic acid bacteria in goat yogurt samples.

Key words: electronic nose, goat yogurt, lactic acid bacteria, multivariate analysis