In this paper we present a case evidence generation architecture, CEG,
that supports case based decision and knowledge modeling. CEG is a bl
ackboard based system that supports case evidence generation and acqui
sition. It can be used to define case decision classes, to model case
knowledge elements, to represent prototypical case features and cases,
to index cases according to decision classes or prototypical case rea
lizations and to model elements of bias or error in case representatio
n. We place CEG in the context of multicriteria decision making and dr
aw architectural components from Machine Learning. We give an overview
of CEG and CEG knowledge components. We demonstrate CEG by means of t
wo 'real world' decision making exemplars drawn from retail sales eval
uation and medical decision making. Finally, we explore architectural
properties and discuss the potential for further research on the subje
ct.