Approximate Reaction Systems Based on Rough Set Theory.

In this work we investigate how Rough Set Theory could be employed to model uncertainty and information incompleteness about a Reaction System. The approach that we propose is inspired by the idea of an abstract scientific experiment: we define the notion of test, which defines an approximation space on the states of a Reaction System, and observation, to represent the interactive process of knowledge building that is typical of complex systems. We then define appropriate notions of reducts and study their characterization in terms of both computational complexity and relationships with standard definitions of reducts in terms of Information Tables.