Extracting compact fuzzy rules based on adaptive data approximation using B-splines

We discuss the importance of making a fuzzy controller human interpretable and give an overview of the existing models and structures for that purpose. We then summarise our approach to designing fuzzy controllers based on the B-spline model by learning. By using an optimal partition algorithm and linguistic modificators like "between", "at most", "at least", etc., the rule base can be reduced to the minimum. This helps to avoid the over-fitting problem and improves the interpretability of the model. We tested the controller on different benchmark problems and achieved a rule compression ratio of up to 71%.