This contribution is focused on the enhancement of the precision for Fuzzy Rule Based Classification Systems by the refinement of the Knowledge Base. Specifically, we make use of a Hierarchical Fuzzy Rule Based Classification System, which consists in the application of a thicker granularity in order to generate the initial Rule Base, and to reinforce those problem subspaces that are specially difficult by means of the application of rules with a higher granularity. Furthermore, we will perform a genetic rule selection process in order to obtain a compact and accurate model. Our experimental results show the goodness of this approach, especially when the number of classes is high, which usually implies a higher difficulty in the separability of the examples. Our conclusions are supported by means of the corresponding statistical tests.
展开▼
