Fuzzy Classifiers

Pages: 185-207

DOI: 10.1142/9789814675017_0005

Abstract:

Fuzzy classifiers as a class of classification systems have witnessed a lot of developments over many years now by various communities. Their main strengths stem from their transparency to human users and their capabilities to handle uncertainty often present in real-world data. Like with other classification systems, the construction of fuzzy classifiers follows the same development lifecycle. During training, in particular fuzzy classification rules are developed and undergo an optimization process before the classifier is tested and deployed. The first part of the present chapter overviews this process in detail and highlights the various optimization techniques dedicated to fuzzy rule-based systems. The second part of the chapter discusses a particular facet of research, that is, online learning of fuzzy classification systems. Throughout the chapter, a good review of the related literature is covered to highlight the various research directions of fuzzy classifiers.

Source: Scopus