Intelligent decision support based on influence diagrams with rough sets

Abstract

Influence diagrams have been widely used as knowledge bases in business and engineering. In conventional influence diagrams, the numerical models of uncertainty are probability distributions associated with chance nodes and value tables for value nodes. However, when imprecise knowledge from large-scaled data set is involved in the systems, the suitability of probability distributions is questioned. This study proposes an alternative numerical model for influence diagrams: rough sets. In the proposed framework, the causal relationships among the nodes and the decision rules are expressed with rough sets from information systems. This study develops rough set-based framework in influence diagrams with an illustrative example.