The Research Of The Prediction Model Of District Macroeconomy And Arithmetic

In macroscopical economic forecasting, because the factors influence economy arevarious, and the economic system is in variational environment, traditional predictionmethod can't get rational and effective result. Then, based on the theory of economicalforecast and warning, this paper uses neural network and evidence theory to research onforecast method.Firstly, to have preliminary research on the early warning index system of economy,the paper designs the index system of forecast and warning adapts to districtmacroeconomy, and divides indexes into determinacy factor and uncertain factor, basedon the main character of macroeconomy, design principle of the index and otherdomestic and international index system of macroeconomy.Secondly, to utilize the neural network that has the characters of self-study,self-adapt, organizing, non-linear and fault-tolerant ability etc, the paper uses BP neuralnetwork that predicts result by historical data for the index of determinacy, to predictuncertain index through relevant expert's appraisal.Finally, on the basis of the success that evidence theory solves uncertain problem inlots of fields, to introduce the evidence theory method and utilize it to merge thepredicting result from . the determinacy and uncertain indexes, the article forecaststhe macroeconomy.. The warning model of district macroeconomy established isrational, proved by application. It indicates the prediction way uses evidence theory andneural network synthetically is effective, accurate and satisfied the goal of districtmacroeconomy's prediction basically.