A Generalized Multigranulation Rough Set Model by Synthesizing Optimistic and Pessimistic Attitude Preferences

Abstract

Attitude preference plays an important role in multigranulation data mining and decision-making. That is, different attitude preferences lead to different results. At present, both optimistic and pessimistic multigranulation rough sets have been studied independently and thoroughly. But, sometimes, a decision-maker’s attitude may vary, which may shift either from an optimistic to pessimistic view of decision-making or from a pessimistic to optimistic view of decision-making. In this paper, we propose a novel multigranulation rough set model, which synthesizes optimistic and pessimistic attitude preferences. Specifically, we put forward methods to evaluate the attitude preferences in four types of decision systems. Two main issues are addressed with regard to attitude preference dependency. The first is concerned with the common attitude preference, while the other relates to the sequence-dependent attitude preference. Finally, we present three types of multigranulation rough set models from the perspective of the different connection methods between optimistic and pessimistic attitude preferences.

Keywords

• rough set theory
• granular computing
• multigranulation rough set
• attitude preference