In this paper, an E-nose was used to detect the adulteration of
pork in minced mutton. Multivariate analysis methods were employed
to explore the performance of the E-nose in classification
of the adulteration. 120 samples were detected and the signals
were analyzed by three feature extraction methods and pattern
recognition techniques. Step-LDA was proved to be the most effective
feature extraction method. It is applicable for the E-nose to detect
the adulteration in mutton based on the Step-LDA using CDA.
MLR, PLS and BPNN were used to predict the pork content in
minced mutton precisely. Three methods showed high capacity
in prediction for content of pork in minced mutton. What’s more,
BPNN was the most effective method for the prediction of pork
content. The E-nose had been proved to be a useful authentication
method for meat adulteration detection for its efficiency and high
accuracy.