Improving Environmental Scanning Systems Using Bayesian Networks

As the environment for companies becomes increasingly volatile, scanning systems gain in importance. We propose a hybrid process model for such systems information gathering and interpretation tasks that combines quantitative information derived from regression analyses and qualitative knowledge from expert interviews. For the latter, we apply Bayesian networks. We derive the need for such a hybrid process model from a literature review. We lay out our model to find a suitable set of business environment indicators to forecast a companys key financials. Deriving lessons learned from a prototype in the industrial sector, we evaluate the utility of our model following the design science research paradigm. We find our model to especially convince in completeness, transparency and transportability when compared with pure mathematical models.