Study On The Information Analysis And Decision Making Method Of Imports And Exports In Foreign Trade Of Chongqing Based On The Support Vector Machine

As a kind of time series, the data of imports and exports in foreign trade has the characteristics of randomicity, complicacy, strong nonlinearity and noisy, especially in local area, its fluctuating and nonlinearity are more evident. So it's difficult to use the traditional timing prediction technology to describe it accurately. At the same time, for some reasons, the data of imports and exports in foreign trade of Chongqing is quite limited in the time dimensionality, and that makes it necessary to use a method with good compatibility in small samples.Support vector machine (SVM) is a new method for pattern recognition based on the statistical learning theory. It is an implementation of structure risk minimization principle in the statistical learning theory. SVM presents a lot of advantages for resolving the small samples, nonlinear and high dimensional pattern recognition, as well as other machine-learning problems such as function fitting.Due to its high fitting precision, perfect generalizability, global optimization and good compatibility with small samples, the Support Vector Machine (SVM) and its mode used into the prediction and decision-making process for the Chongqing foreign trade is proposed in this paper for the first time. Via the analysis of the status quo of Chongqing foreign trade, aim at the problems and data features it has, and set the main objects of analysis. The study in the paper is based on the data in the imports and exports information system of Chongqing foreign trade.Using SVM to do the classification of the samples of the decision system, a simple decision function is obtained, which could forecast the sample's class and act as decision rules. The paper discusses the application of SVM in the foreign trade decision-making in the aspects of class forecast.According to the data features of Chongqing foreign trade, a prediction model for the time series of foreign trade information based on the Support Vector Regression (SVR) is discussed in this paper, while a modeling method for using the forecasting results and Support Vector Classification (SVC) to do the markets classification and selection of decision projects is also proposed. This method was used in the ChongQing Motorcycles export prediction and its market decision-making, which served as a comparison and analysis to its function. The different type of SVM, standardization of the samples, the selection of kernel function and its parameters and its effect on the...