Export Trade Forecasting Of Guangdong Province Based On NN Model

Under the background of economic globalization, export trade plays an important role in economic growth in Guangdong province. Export trade is a complicated nonlinear system influenced by many factors from both domestic and abroad, which make it difficult to be handled by the traditional linear estimation mechanism. However, Neural Network performs well in robustness, fault-tolerance and nonlinear mapping and can be a suitable solution to the estimation and decision-making problem against nonlinear economical system.Firstly, the thesis analyzed the current status of Guangdong export trade industry based on its characters, the thesis then divided all the indices which may affect export trade into 5 groups; Secondly, the thesis applied the NN theory to estimation and modeling for the export trade sum, and multi-factor model and single-time-serial model were built for this. Lastly, against the slow convergence and local optimal solution issues, the thesis provided the NN model based on the genetic algorithm. And found through MATLAB simulation, after optimized by the genetic algorithm, the NN model not only accelerated the convergence of the program, but also achieved higher precision than standard BP Neutral Network.Logical estimation result can assist Guangdong in constituting export trade policy, inducting and organizing export enterprise behavior as well as optimizing export structure. This genetic algorithm based NN model can be regarded as a reference for other economic modeling.