Efficient Context Prediction for Decision Making in Pervasive Health Care Environments: A Case Study

Abstract

Mobile real time decision support systems (RTDSS) find themselves deployed in a highly dynamic environment Decision makers must be assisted taking into account the various time critical requirements Perhaps even more important is the fact that the quality of the support given by the system depends heavily on the knowledge of the current and future contexts of the system A DSS should exhibit inherent proactive behaviour and automatically derive the decision making person (DMP) s needs for specific information from the context that surrounds him/her We propose to run a DSS on top of a middleware that helps the decision maker to contextualise information Mort-over we give a set of requirements that the middleware should fulfil to learn detect and predict patterns in context to optimise the information flow to the decision maker The approach is made concrete and validated in a case study in the domain of medical health care Representative location prediction algorithms are evaluated using an existing dataset