Chapter 4
Rough-Fuzzy Granulation and Pattern Classification
4.1 Introduction
Granular computing refers to computation and operations performed on information granules, that is, clumps of similar objects or points. Granular computing has been changed rapidly from a label to a conceptual and computational paradigm of study that deals with information and knowledge processing. Many researchers [1–3] have used the granular computing models to build efficient computational algorithms that can handle a huge amount of data, information, and knowledge. These models mainly deal with the efficiency, effectiveness, and robustness of using granules such as classes, clusters, subsets, groups, and intervals in problem solving [4].
Granular computing can be studied on the basis of its notions of representation and process. However, the main task to be focused on is to construct and describe information granules by a process called information granulation [5, 6] on which granular computing is oriented. Specifically, granulation is governed by the principles according to which the models should exploit the tolerance for imprecision and employ the coarsest level of granulation, which are consistent with the allowable level of imprecision. Modes of information granulation, in which the granules are crisp, play important roles in a wide variety of approaches and techniques. Although crisp information granulation has a wide range of applications, it has a major blind spot [7]. More particularly, ...
