Abstract: An artificial neural network(ANN)model is constructed to predict the retention time,based on the high performance liquid chromatographic(HPLC)analysis data of seven food additives.The model is created in the structure of feed-forward network with the algorithra of back propagation.It is a double-layer topological network,including one hidden layer.The optimum node number of hidden layer is determined.With these parameter values,the model can be trained accurately,and the training speed are kept in an appropriate range.The network redundancy and the drop of constringency speed are avoided.The simulation results shows that the network can be trained stably by basic BP algorithm,and the predicting results are well in accordance with experimental data.

Key words: artificial neural network (ANN), food additives, high performance liquid chromatographic (HPLC), modeling