Construction Of Sino-US Trade Deficit Model Based On Statistical Analysis

When you study on macro issues of Sino-US trade, need the same data comes from various different countries, the different departments of the country, or from a number of different international organizations。In order to collect the data, the researchers need to check a lot of literature。The guiding ideology of the Sino-US trade system design is to allow researchers to sort out their own data are added to the Sino-US trade system。 Forces gathered many researchers, sharing and use of these data. The data collected by the individual person may exist indicators omissions or missing data, but the use of the China-US trade system can be a good solution to this problem.Due to the needs of various users of the Sino-US trade system under the authority of the provisions of the operation and use of data through the network. Sino-US trade system to select the B/S mode. The core problem of the Sino-US trade is that China has a great surplus in the Sino-US trade, if we can achieve the mathematical model of the US-Sino trade deficit,it can provide theoretical support and policy solution scheme to solve the problems in the Sino-US trade. So we establish Sino-US deficit model, another indicate o design and implementation of Sino-American trade system.In order to achieve the specific model of the US-Sino trade deficit, we use Grey Relational Analysis filter the variables and eliminate variables that has nothing to do with the US-Sino trade deficit; Next, we use the principal component analysis to reduce the Linearly independent variable dimension and achieve impact factor of linear correlation.We take the impact as the independent variables, the US-Sino trade deficit as the dependent variable. We use the multiple linear regression analysis establish the US-Sino trade deficit mathematical model, but the model hypothesis test did not pass, can not be applied in practice. Based on the experience, the US-Sino trade deficit must be affected by the impact factor of linear correlation, so we can think that the mathematical model must exist. On the other hand, we found that the partial derivative of the dependent variable for each impact factor exists by curve regression analysis, so we use Multivariate nonlinear regression analysis establish the estimate model of US-Sino trade deficit.