Modeling Relationships in Data: From Contingency Tables to Fuzzy Multimodels

Publisher Summary The chapter discusses the revealing modeling relationships between variables in exploratory data analysis. These relationships can be modeled through rules, linear or nonlinear functions, and neural networks, to name a few alternatives. The objective of fuzzy multimodels is to deliver an environment assuring a successful interaction among several relational or functional constructs and to allow for their efficient utilization. An interesting and useful scenario arises when the data are governed by relational and one-to-many mappings rather than being confined to purely functional mechanisms. Fuzzy multimodeling introduced in this study is concerned with the design and utilization of families of models rather than single models. Fuzzy multimodels comprise a collection of local models along with the relevant mechanisms of their triggering and aggregating aimed at assuring a suitable interaction among these models. The algorithmic details are laid down and illustrated through several detailed simulation studies.