The Research Based On Artificial Intelligence Of Early-Warning Model For Anti-Dumping On China's Textile Exports

Anti-dumping, regarded as the most common way of protecting the economic in International Trade, has become the focus of many countries. After China's accession 'to the WTO, anti-dumping has become the principal problem of the trade disputes between China and other countries. More and more countries have started their anti-dumping investigations on China's exports of textiles. China is one of the countries which have been investigated and penalized for anti-dumping mostly. For the time being, in fact, many countries have a great strength of textile production, whose cost is also very low. Usually the consequence is that China's textile exports are judged as dumping and the enterprises involved would be forced to surrender the market. Therefore we must establish the effective early-warning method to protect the negative influence from Anti-dumping. This article is started from the current world status of anti-dumping in China, influence and characteristics, it also in detail analyses the internal and external reasons why China always encounters anti-dumping issues frequently, and combines with SOM Neural Net to realize the early-warning model of Anti-dumping by the use of Matlab math-working implement.According to the situation of our country and the international economic circumstances, the main factors, which impacted textile export in the past five to six years, can be analyzed, and the detail data aimed at these factors will be concluded to indicate characteristic indexes of the neural net. The Neural tool box in Matlab will imitate the role of SOM neural net to help us to fulfill the classification of these indexes. Compared the result with the reality, the link between the category of indexes and the critical extent of Anti-dumping can be generated so as to establish the early-warning system to foresee the Anti-dumping.The research has approved that the Anti-dumping prediction method by use of SOM neural net is feasible, and has reached the accurate rate. It is the system based on SOM neural net that lets the Anti-dumping early-warning process more and more convenient, which do not take the Anti-dumping mechanism into consideration, just fill some indexes in Matlab. After that the system will make a judgment via the critical extent of Anti-dumping situation, and we can carry out macro-control when textile exported, thus do the best to simplify the analysis of Anti-dumping so that the prediction can be easier and agiler.