Decision support systems (DSS) are important tool in accurate and realistic decision making process. To achieve consistent results irrespective of the user's decision making skills, however, there is the need for an adequate model of the decision makers' actions in the context of human-computer interaction (HCI). This task requires that users' actions be related to the decision making processes in a manner that is transparent and that is appropriate to individual decision maker's skill. I will discuss our efforts at developing an HCI model for achieving these in the context of a medical DSS environment. We used DSS for suicide susceptibility prediction as our case study.
Issues relating to the requirements of an HCI in DSS domain and the theories that underlay our human-computer interaction model will be discussed. Our proposed approach, which exploits the integration of fuzzy constraint reasoning and ontology as well as its realisation using formal methods, will be presented.
I will draw the conclusion that the combination of the processing power of constraint based reasoning, the rigour of formal methods, the ability of the fuzzy logic paradigm to model imprecision, vagueness and uncertainty as well as the ability of ontology modelling for simplifying the representation of concepts with one connotation but multiple semantic implications, results in a powerful HCI modelling framework.