Aplication Of GMDH Network With PSO Algorithm Optimization In The Prediction Of Import And Export Trade

Since our country initiate the reform and opening-up policy, with the rapiddevelopment of economy, Our country foreign trade develop rapidly. Total foreigntrade volume increase at a double-digit growth. From1978to2012, our country’stotal import and export increases from20.64billion to3.87trillion, annual import andexport increase rate reach15.4%. As the economic center of china, Shenzhen city isthe first to be designated as a special economic zone and a showcase of China’sopening up. It’s import and export increases more rapidly. According the economytheory, because of high volubility and many factors, import and export cannotpredict precisely.On the basis of extract indicators, the article built Shenzhen importand export trade predict model base of GMDH network.The article first introduces the situation of Shenzhen import and output, then itbriefly introduces Shenzhen,import and export character, import and export structureand main import and export countries and regions. Then,it constructs GMDHnetwork prediction model based on PSO optimization. On the basis of introduce PSOand GMDH, it analyses the PSO optimization GMDH model, method and flow. Onthe basis of extracting eight indicators which is relevant with Shenzhen import andexport from inner economic environment, strategy environment indicator andinternational economic environment. It test the indicator correlation andPCA.Lastly,Shenzhen value of import and export is used to test the prediction ofGMDH network and PSO optimization GMDH network. The two network is used topredict Shenzhen value of import and export. The simulation experiment shows that,compare with the signal GMDH network, PSO optimization GMDH network canpredict precisely.The study of the article shows that GMDH network can precisely predict thedynamic change of Shenzhen import and export trade with the advantage ofadaptively construct. It provides a new method for import and export trade prediction.