Rough Neurocomputing: A Survey of Basic Models of Neurocomputation

This article presents a survey of models of rough neurocomputing that have their roots in rough set theory. Historically, rough neurocomputing has three main threads: training set production, calculus of granules, and interval analysis. This form of neurocomputing gains its inspiration from the work of Pawlak on rough set philosophy as a basis for machine learning and from work on data mining and pattern recognition by Swiniarski and others in the early 1990s. This work has led to a variety of new rough neurocomputing computational models that are briefly presented in this article. The contribution of this article is a survey of representative approaches to rough neurocomputing.