INFORMATION GRANULES
Władysław Homenda, Witold Pedrycz · Pattern Recognition · 2018
This chapter focuses on the fundamental constructs of granular computing, which are of importance when processing information granules and formulating some general ways in which the results are developed and interpreted. The concepts and algorithms presented in the chapter are of direct usage in various problems of pattern recognition and data quality. The chapter emphasizes how information granules are developed on the basis of some well-defined and intuitively supported objectives. The concept of the granular mapping delivers a generalization of commonly encountered numeric mappings no matter what their structure is. In this sense, this conceptualization in the format of granular mappings offers an interesting and practically convincing direction. The constructs of this nature are valid for any formalism of information granules. In system modeling, granular models arise in a hierarchical way by raising a level of information granularity at successive layers of the architecture. The chapter elaborates the essence of the formation of such models.