Fuzzy Rough Set Theory‐Based Feature Selection
Tanmoy Som, Shivam Shreevastava, Anoop Kumar Tiwari, Shivani Singh · 2020
A variety of techniques for attribute reduction in fuzzy rough set (FRS) environment are widely discussed by many researchers. This chapter considers three types of decision systems, i.e. supervised (all the class labels are available), semisupervised (some of the class labels are available), and unsupervised (all the class labels are missing). It examines attribute reduction of decision systems with missing attribute values. Various application areas induced from FRS-based approaches for attribute reduction include especially classification problems. The chapter presents some limitations of FRS-based techniques and suggests measures to overcome them. It also discusses some preliminaries on the rough set, fuzzy set, and FRS. The chapter provides a detailed survey on fuzzy rough-assisted attribute reduction for supervised decision systems. It presents various techniques for feature selection of semisupervised and unsupervised decision systems. Finally, the application of fuzzy rough set theory (FRST) in decision-making problems is discussed.