Research On Support Vector Regression And It's Application To The Food Security

Statistical Learning Theory ( SLT for short) is a small-sample statistics theory brought forward by Mr. Vapnik et al. Support Vector Machine ( SVM ) is a new machine learning method based on Statistical Learning Theory. It has become one of the most important achievements in machine study field in the last ten years, Reggression is one of the most important application in SVM.At the beginning this article introduces related important conclusions of Statistical Learning Theory and support vector reggression, a model about weighting support vecor reggression is proposed in order to solve the different noise level of samples; we know Kernel function and parameter selection are very important in the support vector reggression, they are disccused and given some selection methods based on the information given samples.A support vector reggression model based on the Rough sets Theory is presented in the paper, Utilizing the advantages of RS theory in processing large data and eliminating redundant information has decreased SVR training data and overcome the disadvantages of very large data and slow processig speed caused by SVR approach.At last in the paper, support vector reggression and it's kernel method are applied to food security , giving a kind of comprehensive evaluation model about food consume level and production level of thirty one Provinces and Cities in our country based on kernel principal component analysis , at the same time, comparing food security about our country with some other countries. And giving a kind of economic forecasting model based on SVR, Comparing with the method of BP neural network and ARIMA model. It is denoted that the SVR and its Kernel methond has advantage and its feasibility in the Food security.