Get access

Evolving fuzzy systems for data streams: a survey



Evolving fuzzy systems (EFSs) can be regarded as intelligent systems based on fuzzy rule-based or neuro-fuzzy models with the ability to learn continuously and to gradually develop with the objective of enhancing their performance. Such systems learn in online mode by analyzing incoming samples, and adjusting both structure and parameters. The objective of this chapter is to present a brief overview of some early as well as recent EFSs by focusing on their architecture, design algorithms along with the merits and demerits, and various applications. © 2011 John Wiley & Sons, Inc. WIREs Data Mining Knowl Discov 2011 1 461–476 DOI: 10.1002/widm.42