Rough-Granular Computing for Relational Data

Rough set theory and granular computing have widely been applied in data mining. They have been used separately as well as a combined approach called rough-granular computing. The usefulness of this approach in data mining is the driving force for employing it to improve processing of relational data. This chapter introduces three rough-granular approaches dedicated to handle complex data such as relational one. Each of them processes relational data as granules and use the tolerance rough set model to deal with possible uncertainty in data. The chapter also compares the three approaches in terms of construction of information system, information granules, and approximation spaces.