Modelling of rough-fuzzy classifier
Jiří Křupka, Pavel Jirava · 2008
Abstract: The paper reflects the trend of the past years which is based on the diffusion of various traditional approaches and methods to the way of tackling new problems, in this case to the classification. Two components of the computational intelligence are applied in a classification model. It means rough and fuzzy sets on the basis of which the data classification hybrid model is proposed. In the second part the current knowledge in the investigated field are summarized and briefly explained. The algorithm for uncertain data operations and conditioned rules generation is introduced too. We have brought in an original toolbox called RSTbox for data processing and automatic rules generation. Proposed model and toolbox are carried out in MATLAB, tested on more data files, and compared to others, already known classification methods.