Data-driven decision making: new opportunities for DSS in data stream contexts

Abstract

Traditionally, Decision Support Systems (DSS) data were stored statically and persistently in a database. Increasing volume and intensity of information and data streams create new opportunities and challenges for DSS experts, data scientists, and decision makers. Novel data stream contexts require that we move beyond static DSS modelling techniques to support data-driven decisionmaking. Implementing incremental and/or adaptive algorithms may help to solve some of the challenges arising from data streams. This research investigates the use of these algorithms to better understand how their performance compares with more traditional approaches. We show that an adaptive DSS engine has the potential to identify errors and improve the accuracy of the model. We briefly identify how this approach could be applied to unexpected highly uncertain decision scenarios. Future research considers new opportunities to pursue a multidisciplinary approach to adaptive DSS design, development, and implementation leveraging emerging machine learning techniques in tackling complex decision problems.