Inference Systems by Using Ordinal Sums and Genetic Algorithms

In this work a new approach developed by using Ordinal Sums to apply general t-norms to inference systems of different neuro-fuzzy systems is proposed. A Genetic Algorithm based strategie to search the best t-norms and/or t-conorms from data is adopted. By using the approach two known neuro-fuzzy systems, that are the Fuzzy Basis Function Network and the Fuzzy Relation Neural Network models are compared. Several experiments on synthetic and benchmark data using different parametric and non-parametric t-norms and t-comorms are made.