ABSTRACT
This chapter summarizes some of the classical work on concept identification and how this work relates to the more research on learning from cases and learning by exploration. In the research on concept identification, subjects were typically put into well-specified closed-world situations. Because the experimenters had completely defined these closed-world situations, they also had complete knowledge about all aspects of these artificial situations. A case is a specific situation, which is described in terms of a problem and its solution. It represents knowledge that is at an operational level and that is tied to a context. The cases and concepts that occur in real life can be quite complex. The research by Rosch on natural concepts has shown that although objects could be classified into one of several hierarchically structured categories humans tend to use the basic level categories. Various computational models have been developed to describe the cognitive processes that occur in learning from cases.
