Decision Automation for Predictive Analysis Models

A methodology for developing an identification map for economic predictive systems that can be analyzed via Bayesian belief networks is proposed in this paper. The methodology describes how to automatically design a diagnostic decision tree from a Bayesian belief network used for business intelligence predictive analysis. In particular the methodology makes use of econometric characteristic functions, since they are mathematical models used for predictive analysis. We used the Vesely-Fussell measure of importance as the cornerstone of our methodology, because it provides an accurate measure of components' relevance from a diagnosis perspective. The essence of this research and paper is to apply diagnostic analysis to econometric systems, thus enabling the identification of root causes of changes in econometric objectives using reliability engineering theory. To perform diagnostic analysis on econometric predictive systems analysis, we demonstrate how to construct a diagnostic decision tree, which we previously developed for diagnosing hardware/software systems.