A Study of Tongue Manifestation Feature Fusion Using Rough set and SVM Probability
基于粗糙集与SVM概率输出的中医舌象特征融合方法研究

Multi-feature fusion of tongue manifestation is studied, using the rough set and SVM probability output.The samples are divided into two new groups: defined group, and uncertain group. The support vector machine classifier that produces probability is trained to the re-classified samples. The method is able to produce the probability needed for characterizing each group, and is desirable for studying the fusion of four diagnosis methods.