Artificial Neural Network Model In The Regional Macroeconomic Forecasts And Methods

Neural network is a class of new and developing frontier science.It has been more than 50 years since this theory has been firstly put forth ml 943 .During these years,It grows rapidly and includes many flelds.Especial in following flelds:Identiflcation of model,Identiflcation of system,Intellectualized control,lmage managing process,Diagnosis of malfuncation,Diagnosis of disease and Economic forecasting. In this paper,we use BP method of artificial neural network to forecast practical economic system..Using artificial neural network to forecast regional economy, at the same time to deeply investigate internal mechanism of artificial neural network,and to improve the model that we had ever used before. Then we put forth some new effective method of neural network. Main work are summarized as follows: ? Summarizing feed-forward neural network, proposing some new effective method of neural network which include improvement of model algorithm and structure of network. Using feed-forward artificial neural network and BP algorithm to forecast increase of three types of production,general nation production and other kinds of economic index.Analysing forecasting results.Proposing some reasonable suggestion for development of Liaoning economy. ? Establishing reasonable road-net is important content to road-net project,is important target aim to macro control to road construction.in this paper using artificial neural network model to forecast gross milage of Henan province,and achieving reasonable,effective forecasting results. Summarized several new neural network algorithms including Changing scale method,Common gradient method,Least different multiplication~ in order to improve researching efficiency and stability of multi-layer feed- forward neural network and overcome some limitation of BP algorithms.