State-Based Reconstructability Modeling for Decision Analysis

Michael S. Johnson and Martin Zwick.

Proceedings of The WorldCongress of the Systems Sciences and ISSS 2000, Allen, J.K. and Wilby, J.M. eds., Toronto, Canada: International Society for the Systems Sciences.

Abstract

Reconstructability analysis (RA) is a method for detecting and analyzing the structure of multivariate categorical data. Jones and his colleagues extended the original variable-based formulation of RA to encompass models defined in terms of system states (Jones 1982; Jones 1985; Jones 1985; Jones 1986; Jones 1989). In this paper, we demonstrate that Jones' previous work comprises two separable ideas: the "g to k" transformation and state-based modeling. We relate the concept of state-based modeling to established variable-based RA methods (Klir 1985; Krippendorff 1986), and demonstrate that state-based modeling, when applied to event and decision tree models, is a valuable adjunct to the variable-based sensitivity analyses commonly employed in risk and decision modeling. Examples are provided to illustrate the approach, and issues associated with the interpretation of state-based sensitivity analyses are explored.