Adaptive knowledge discovery in expert systems

Decision support systems (DSS) are extremely beneficial in many data science fields. The increasing amount of data complicates the process of keeping the overview. For example, in the medical environment advanced DSS are available to assist medical experts while diagnose and treatment process of patients. The paper discusses two major issues which still remain challenging. First, the comprehension of the output from prediction models. Experts need to understand the decision-making process in detail to consider a suggested diagnose or treatment. Second, to improve the decision making, the expert's knowledge needs to flow back by evaluating the outcome. By improving through evaluation, the algorithm adapts over time. One of the major challenges in every step is to be transparent. The paper describes the process model for adaptive knowledge discovery in expert systems by addressing the aspects of data quality, knowledge extraction and recommendation.