ABSTRACT
Computational intelligence is a research area oriented to model different aspects of intelligence embracing methodologies such as neural networks, fuzzy systems, and evolutionary algorithms. This chapter presents the imbalanced dataset problem and describes the oversampling, undersampling, and hybrid methods used in this work. It also presents basic concepts of Fuzzy Rule-Based Systems and Fuzzy Rule-Based Classification Systems. The chapter describes the Genetic Fuzzy Systems. It also describes the method proposed for preprocessing the imbalanced dataset and learning fuzzy classification rules using the Iterative Rule Learning approach and Multi-Objective Evolutionary Algorithms (MOEA). Several real-world problems are characterized by imbalanced learning data. MOEA's are used in problems with multiple conflicting objectives, where the improvement of an objective leads to the deterioration of the others. Recently, MOEAs have been adopted as a more suitable optimization technique to generate fuzzy systems.
