Based On Bp Neural Network Of Environmental Effect Forecast China's Trade

In the past few decades, China was implementing the transformation of theeconomic development strategy from close to open, and started to implement theliberalization of trade policy. In recent years, in the great swing of economicglobalization and the international division of labor of specialization, the export tradescale of our country presents a trend of expansion, and the resulting problems ofenvironment relationship have attractted great attention from the scholars andgovernment. If we blindly pursue the trade liberalization without considering thesustainable development, the trade expansion may bring great negative influence tothe environment, and even the environmental costs will greatly exceed the economicinterests. Therefore, in order to provide reliable basis for the government to makereasonable environment and trade decisions, and to realize the healthy and sustainabledevelopment of the trade liberalization of our country, it is necessary to do accurateprediction of inherent rules and developing trend of the environmental pollutioncaused by trade.In the practical ecomonic system, the relationship between the tradeliberalization and the environment is not the simple one from trade liberalization toincreased output and then to increased pollution. Trade brings influence to theenvironment through different mechanisms. In addition to the scale effect mentionedbefore, there are structure effect and technology effect. The influence that the tradeliberalization brings to the environment is a very complex problem. From the theoryand model aspects, this article will firstly analyses how the trade influence ourcountry’s environment through these three effects. Theoretical analysis shows that theeffect of environment caused by trade scale is negative. The structure effect of tradeproducts should depend on the comparison of the pollution intensity between theproduction expansion department and the production contraction department. Thetechnology effect of environment caused by trade restrains the pollution level, and the effect is positive. A later established small model of open economy confirms theabove theory mathematically. And then, this article introduces the basic theory ofartificial neural network, stresses fully discussion on the basic principle, learningalgorithm and the shortage and improvement of the BP neural network, and designsthe network combined with the basic theory and the specific problems researched inthis paper. Finally, combined with the above theory and the model analysis, weestablish a forecast index system between the trade and environment, realize thesimulation of the model system by using the best network structure and parametersselected by trial and error in the MATLAB software, and then conclude the predictedvalue of the environmental pollution index of our country.The complicated nonlinear relationship between trade and environment rulesout the applicability of a number of traditional predictive methods on this problem.But the adaptive learning and fault-tolerance ability of BP neural network display agreat disadvantage at this moment. The process of the positive research shows that theBP neural network provides a new and effective way for the prediction of tradeenvironment effect. Although the BP neural network itself still exists certaininsufficient and defects, it can reach the expected accuracy target by selecting theoptimal learning algorithm and debugging repeatedly.