Study on early-warning system of pork price based on BP neural networks

Abstract

In order to predict pork prices and its trend, early-warning model based on the BP neural networks was established to protect the interest of producers of pigs and consumers. The programming of C# and MATLAB 2008b was mixed in implementation early-warning system of pork prices. C# realized man-machine interface of early-warning system and database operation, the training and simulation data of neural network self-learning can be added or imported to database. When the neural network was trained and simulated to predict pork price, the BP neutral network model which was designed by MATLAB was called in the background, and certainly the performance, training status, and regression chart of neural network model can be illustrated by MATLAB. Combined with regulation plan about preventing pork price fell excessively which was jointly issued by relevant state departments in 2009, when the pork price fluctuations is over 15% or ratio of pork price and grain price is less than 6,early-warning system will send out early-warning information. If price index parameters of other animal can be determined, the pork price early-warning would be expanded into other animal's price early-warning. Early-warning of animal price can provide the basis for government administrators and guide animal productions, the economic loss will be avoidable by reducing price fluctuation.