The Design Of Grey Neural Network Forecast System And Its Application

The grey system theory has provided a new academic and technical support for the data processing of small sample and poor information in data mining.Grey forecast theory is an important branch of the grey system theory,especially,the GM(1,1) model,as the core of the grey forecast system,has been widely used.For the unique in processing the data which has the characteristics of poor information and small sample,the general researchers deeply like to use them. The BP neural network with its distinct network structure and good function fitting effect, has been widely used in the prediction, optimization, management, etc.In order to give better play to grey forecast system and BP neural network’s advantages,this paper will explore their combining forecast method to set up the new forecast method,develop new forecast system and explore new application way.This paper systematically has analyzed the research status of grey system and BP neural network forecast,looked for the combining site of two forecast methods and improved the performance of traditional grey forecast methods.It constructs a BB-BP-GM(1,1) model through combining the improved grey forecast model and BP neural network,and develope the forecast application system based on the BB-BP-GM(1,1) model,further improve the accuracy of forecasting. At the same time,it thoroughly research the combining site of non-spacing interval grey model and BP neural network to construct the improved non-spacing interval grey BP neural network forecast system.In this paper,the forecast system,based on BB-BP-GM(1,1) model,is applied to the traffic accident forecast of Heber province,through compared with the forecast results of the improved non-spacing interval grey BP neural network forecast system,it confirm the combining models have better performance than the traditional models,and the establishment and application of combining forecast system have certain practical significance.