Rough set theory can effectively be applied in attribute reduction and rule extraction. To resolve the problems of the reduction of input variables and the difficulty of fuzzy rule extraction etc in the fuzzy system designing process,a method to design fuzzy system is put forward by combining rough sets with fuzzy sets together.
Concept lattice is a hierarchical structure of concepts defined by a binary relation, as a tool for data analysis and rule extraction, which has inspired intensive research in both theory and applications.
Furthermore, let the membership function as the base, we put . forward another algorithm-LBR and after the comparison of LBR and LEM1 which advanced by Kansas University in American, we get an ameliorative algorithm named LEM3. LEM3 is the variable-precision modal of LEM1 factually and the notion of variable-precision can be founded on the measurement of dependent degree.
In chapter 4, based on rough membership value a vital algorithm of getting rules is presented, and the practical example shows that this algorithm has a content result. This has showed an aspect of the practice of the rough sets theory.
A rule extraction tech nique is then applied in order to extract explicit knowledge from the trained neural networks and represent it in the form of fuzzy if-then rules.
Our proposed network adapts from an initial analog phase, which has good training behavior, to a discrete phase for automatic rule extraction.
First, Radial Basis Function Neural Networks (RBFNN) learning techniques are explored, as is usual in the literature, since the local nature of this paradigm makes it a suitable platform for performing rule extraction.
By using support vectors from a learned SVM it is possible in our approach to use any standard Radial Basis Function (RBF) learning technique for the rule extraction, whilst avoiding the overlapping between classes problem.
Linguistic Rule Extraction From a Simplified RBF Neural Network