From knowledge-based to data-driven fuzzy modeling

Abstract

This paper elaborates on a development in (applied) fuzzy logic that has taken place in the last couple of decades, namely, the complementation or even replacement of the traditional knowledge-based approach to fuzzy rule-based systems design by a data-driven one. It is argued that the classical rule-based modeling paradigm is actually more amenable to the knowledge-based approach, for which it was originally conceived, and less so to data-driven model design. An important reason that prevents fuzzy (rule-based) systems from being leveraged in large-scale applications is the flat structure of rule bases, along with the local nature of fuzzy rules and their limited ability to express complex dependencies between variables. As an alternative approach to fuzzy systems modeling, we advocate so-called fuzzy pattern trees. Because of its hierarchical, modular structure and the use of different types of (nonlinear) aggregation operators, a fuzzy pattern tree has the ability to represent functional dependencies in a more flexible and more compact way.