Application of ANN in food safety early warning

In recent years, frequent occurrence of food safety crisis has seriously affected people's health, which causes widespread concern around the world. To effectively track and trace food has become an extremely urgent global issue. Early warning of food safety can prevent food safety crisis. However, there is still very few automatic tracking systems for the entire food supply chain. In the paper we propose a data mining technique to predict food quality using back-propagation (BP) neural network. Some prediction errors could occur when predicted data are near threshold values. To reduce errors, data near the threshold values are selected to train our system. Special care of threshold values and performance of our proposed algorithm are discussed in the paper.