How MDL can lead to testable predictions in human classification
Dr Emmanuel Pothos, Department of Psychology, University of Wales (Bangor)

Abstract: We adopt the interpretation of rationality according to which an organism's behavior is rational if it is optimally adapted to its environment (Anderson, 1990, 1991a, 1991b). Rationality, according to this view, often implies mechanisms that are as informationally efficient as possible. We interpret the problem of dividing objects into categories as one of data compression, to define a framework whereby the best classification on a set of items is the one that maximally compresses the description of the data set. In this approach, categorization is seen as a compression maximization process and empirical results supporting our hypothesis are presented. This work can be viewed as a formal rendering of the basic categories proposal by Rosch (e.g. Rosch & Mervis, 1975).

(Work with Professor Nick Chater, Department of Psychology, University of Warwick)

This seminar was held at the Department of Computer Science, Royal Holloway, University of London on 21 January 1999.