Research on attribute interval optimization method for segmentation based SVM and the Decision Tree Learning

In recent years, statistical learning theory ( for short SLT)and support vector machine(for short SVM) has become an international field of machine learning new hotspot. Decision tree classification learning algorithm is one of the most widely used and very practical inductive reasoning method, In machine learning, data mining, signal processing, intelligent control, artificial intelligence area has a very important theoretical meaning and practical value. It has been successfully applied Wide range of fields from learning medical diagnosis to assessment the credit risk of learning loan applications. Based on Decision Tree Learning and SVM has obvious advantages in their respective, the integration of decision tree and SVM learning will solve the problems that exist between the two effective ways to improve learning ability and decision tree classification performance, give full play to the advantages of both to make up for both insufficient to resolve the two difficult to solve long-standing problems. This paper based on support vector machine classification surface model and support vector point distribution, with the decision tree Classification surface and support vector machine Classification surface effective approximation as the basic starting point, explore new ways on the fusion of decision tree and SVM. Focused study based on support vector points and categories surface shape characteristics´ attribute interval optimization method for segmentation in order to achieve the performance of decision tree optimization.