The export of aquatic products plays a vital role the foreign trade of China. Giving the increasingly grown problems brought out by the export trade risks of the aquatic products, the thesis aims at study early-warning system of export trade risks of the aquatic products to provide defending theory and measures for aquatic products related enterprises and responsible bureaus.Focusing on the characteristics and problems of the export trade risks of the aquatic products in China, the behaviors of marketing risks and policy risks presented in the course of production, process and sale are analyzed. According to related principles, corresponding a framework of early-warning indicators is built with export value as the warming situation indicator and average export price, national total yield the trade gaps of China-USA and China-EU, the marketing shares of aquatic products in USA, the marketing shares of aquatic products in Korea, the marketing shares of aquatic products in Japan as the wanning omen indicators. The method of correlation analysis of time difference is applied to filter the indicators and determine their precursor length; section analysis is used to fix the warn degree bound and its warn limit line. The warning limits are as follows: none warnings L≥18.15, light warnings 15.35≤L< 18.15, middle warnings12.54≤L<15.35and serious warnings L<12.54An early-warning model for trade risks based on support vector machine was developed to study the risks export trade of aquatic products. Employing Crossing Validation selecting anova as the kernel function and indicators of warning situation and warning omen as variables, predictions were conducted the by the model on software of WinSVM with data from 1994 to 2004. It was showed that values of insert and extrapolate examinations both were 100 percent indicating satisfying warning performance. |