Advances in fuzzy systems and networks
ALEXANDER E. GEGOV · 2012
This plenary paper describes two novel fuzzy modelling methodologies for complex processes that are characterised by uncertainty, dimensionality and structure. The first methodology is based on rule base compression in fuzzy systems whereby model efficiency is improved without loss of model accuracy. The second methodology is based on fuzzy networks with modular rule bases whereby model transparency is improved at minimal loss of model accuracy. The two methodologies are validated comparatively on several case studies. The comparison shows the overall superiority of these methodologies to the current methodologies of rule base reduction in fuzzy systems and fuzzy networks with chained rule bases.