A New Fuzzy Support Vector Method

In this dissertation, For data with numerical condition attributes and decision attributes, a new kind of fuzzy support vector machine based on support vector machine theory is proposed. The designing fundamentals and method of computation and realization are given. The experiment results show, the new type of fuzzy support vector machine in this dissertation is more effective. The characteristic of the new fuzzy support vector machine is that, for the data with numerical decision attributes, this algorithm fuzzifying the samples to some classes based on actual problem first, then training them; for new sample, this algorithm not simply dope the corresponding decision value, but give the corresponding class and the membership degree belongs to it based on the actual circs and its condition attributes. This decision result meets the actual circs better, more external and easier to be understood. Emulational experimental result shows that this new fuzzy support vector machine method not only has higher classified accuracy, but also has stronger test capability for the membership degree. This new method optimizes the classified result of support vector machine, enhance the intelligent level of support vector machine.