A Multiclassification Model Based on FSVMs
Bao Qing Hu, Jing Yang, Jinrong He · 2005
Support vector machines (SVMs) proposed by Vapnik are the new method for small sample learning and are widely used in pattern classification and regression estimation. In multiclassfication there exist unclassifiable regions. In other words, some data are unclassifiable. This paper connects fuzzy membership with SVM to solve this problem, and gives a new classification model based on fuzzy support vector machines (FSVMs).