Fuzzy support vector machines for biomedical data analysis

The generalization ability of SVMs is unreliable when the user selects SVMs randomly to classify data examples. The paper proposes a fuzzy system called fuzzy support vector machines (FSVMs) to deal with the problem. Margin values from three different SVMs are fuzzified, combining with the accuracy information of each SVM. The final decision is determined based on all of the SVMs. Experimental results show that the proposed fuzzy SVMs are more stable and reliable than randomly selected SVMs.