Dynamic fuzzy-reasoning-based function estimator

A new fuzzy-reasoning-based system has been developed based on a case base with dynamic fuzzy rules for interpolation and extrapolation in a higher-dimensional space. It retains the strengths of fuzzy systems but circumvents the problems of knowledge acquisition and updating of rules and membership functions. For different applications, we just change the case base. Like an apprentice, the dynamic fuzzy-reasoning-based function estimator (DFFE) would become more experienced, by updating its case base. Three sets of real-life data were used to validate the approach. The DFFE predictions were also compared to results reported for two other fuzzy reasoning techniques. Its performance is very encouraging.