A new influence diagrams approach based on grey rough set for interval valued information system

Abstract

Influence diagram is an effective mathematical representation of decision situations, in the traditional influence diagrams, the dependencies between nodes of uncertainty and the value table of value nodes are constructed through probability models, they get widely applied in intelligent decision analysis. But information system (IS) in our lives are very complicated, such as grey information data, they are imprecise or nondeterministic, knowledge hidden in such IS mostly is imprecise, and how can we express approximate knowledge is the key to conventional influence diagram models. Grey modeling approach and rough sets theory have proved to be useful in data analysis of uncertain information systems. Therefore, this work presents grey rough influence diagram model, it is a combination of grey rough sets and influence diagrams, in which dependencies are computed based on approximate measurement instead of probability distribution in traditional influence diagram. This new approach provides a more comprehensive scheme for rules expression and data analysis with respect to nondeterministic systems.