Rule extraction based on rough fuzzy sets in fuzzy information systems

Abstract

Rough fuzzy sets are an effective mathematical analysis tool to deal with vagueness and uncertainty in the area of machine learning and decision analysis. Fuzzy information systems and fuzzy objective information systems exit in many applications and knowledge reduction in them can't be implemented by reduction methods in Pawlak information systems. Therefore, this paper provides a model for rule extraction in fuzzy information systems and fuzzy objective information systems. This approach uses inclusion degree to propose and represent a new and low computation complexity way for knowledge discovery and rough fuzzy concept classifier in fuzzy information systems and fuzzy objective information systems. Also, an illustration example in the construction sector is presented. This approach is a generalization of rough set model for fuzzy information system. Theory and method of attribute reduction under inclusion degree are suggested in this paper. This approach extends the classical rough set theory from complete information to fuzzy information system. This proposed model is useful for rule extraction in fuzzy information systems and fuzzy objective information systems to figure our knowledge reduction in fuzzy decision systems. © 2011 Springer-Verlag Berlin Heidelberg.