Degrees of Freedom and Advantages of Different Rule-Based Fuzzy Systems

Rule-based fuzzy systems are gaining increasing importance for classification in many fields of application. Various degrees of freedom for the construction of rule-based fuzzy models are analyzed here, comprising fuzzy sets shape, different types of norms, axes rotation, and weights for antecedents and consequents of each rule and for different rules. Results of application on an example dataset are discussed in terms of classification performances, taking into account interpretability at the same time.