An electronic nosed based nondestructive method was established for detecting frozen pork quality as well as determining whether the pork had been refrigerated at 4 ℃ for a long time before frozen. 5 cm thick square frozen pork was added into a 100 mL beaker and sealed by a membrane; PEN3 electronic nose was used to detect the gas in the upper beaker. The data were analyzed by principal component analysis and variance analysis to understand the distribution of multiple data. Secondly, the prediction model of frozen pork and whether it has been refrigerated at 4 ℃ for a long time before the frozen storage was established. There are significant differences of electronic nose data of frozen pork at different storage as well as whether it had been refrigerated at 4 ℃ for a long time. Both linear discriminant analysis and neural network algorithm based on multi-layer perceptron can establish a successful prediction model. The predictive model based on neural network algorithm has better prediction. The detection method is simple, fast and efficient and nondestructive. It lays a technical foundation in order to enhance the supervision and management for frozen meat.
