Extraction of Interpretable Fuzzy Rules from Data



The building of fuzzy or neuro-fuzzy systems is the most promising and appropriate technique for knowledge extraction from databases in the industry, medical applications or in business. Most of these systems are time-varying in the sense that their characteristics are changing with time and so must be the fuzzy systems describing them. On-line building means the insertion, deletion and/or modification of fuzzy rules as the data arrives. When a model is developed using automated data-driven techniques some degree of redundancy and unnecessary complexity cannot be easily avoided. To achieve the goal of complexity reduction and interpretability clustering techniques are combined with rule base simplification and reduction methods.


TS fuzzy models, eTS fuzzy models, fuzzy clustering, interpretability, rule base simplification and reduction.


Soft Computing


Fuzziness in Finland Conference - FiF '04 and Course on Soft Computing in Business and Economics, May 2004

Cited by

No citations found