Case-based decision theory: From the choice of actions to reasoning about theories

Abstract

In the 1990s, David Schmeidler and Itzhak Gilboa initiated the study of decision- making under uncertainty in a completely new framework, without states but with data sets as the information on which to build choice behavior. While the first formulations of case-based decision theory (CBDT) aimed at applications in economic decision-making, this theory which takes data as a primitive concept provides an alternative foundation for deriving beliefs and driving the choice of predictions. This opened a new perspective on old questions in statistics and artificial intelligence. In this review, we summarize these developments in CBDT and highlight the immensely innovative nature of David Schmeidler's academic work.