Research On The Influence Factors Of China's Export Volume Toward The 21th Century Maritime Silk-Road Country

The strategy of "maritime Silk Road" in twenty-first Century is a major strategy for deepening regional trade cooperation.Strengthen our country along the way along the country's trade,by the help of the new normal in the low growth environment to get new trade growth points.This paper uses the classic gravity model on twenty-first Century "maritime Silk Road" along the 14 national panel data analysis and Research on regression,regression results show: China and along the country's export trade between the two countries and GDP,port distance,trade intensity,liner effect,and the connectivity index were.Effect of export trade on the GDP and the distance variables are more sensitive to the performance.This paper also studied the gravity model simulation and region of gravity model,found that two countries export potential is the largest in Egypt and Malaysia;exclude Iran due to geopolitical trade leads to the fluctuation of the situation,the export potential of two countries is the smallest in Vietnam and the United Arab Emirates;relatively small countries of ASEAN trade potential the Middle East and South Asia;trade potential 2015 research object increased significantly;we can also found the close degree of GDP,trade and distance of Middle East countries compared to the Middle East countries are more sensitive to the ASEAN countries on the liner Unicom status is more sensitive.On this basis,this paper as the input parameter of BP neural network was obtained to calculate reference model,and genetic algorithm to optimize BP neural network,the neural network parameters to search the global optimum,the trade volume of export country for the next five years was predicted,test results show that the parameters and weights after the threshold of the BP neural network optimized by genetic algorithm can converge to the target range of error in a more efficient,more smooth and error curve can effectively search the global optimal full-time and threshold,and therefore makes the various characteristics of the GA-BP neural network has the fitting and the forecast effect is more stable and reliable.