A Rational Account of Categorization Based on Information Theory

TL;DR

This paper introduces an information-theoretic rational model of categorization that effectively explains human categorization behavior, matching or surpassing existing models in accounting for classic experimental findings.

Contribution

It proposes a novel categorization theory based on rational analysis and information theory, demonstrating its effectiveness against established models.

Findings

The model explains key categorization experiment results as well as or better than existing models.

It accounts for human categorization behavior using an information-theoretic approach.

The model outperforms traditional cue, context, and hierarchical models in explaining experimental data.

Abstract

We present a new theory of categorization based on an information-theoretic rational analysis. To evaluate this theory, we investigate how well it can account for key findings from classic categorization experiments conducted by Hayes-Roth and Hayes-Roth (1977), Medin and Schaffer (1978), and Smith and Minda (1998). We find that it explains the human categorization behavior as well as (or better) than the independent cue and context models (Medin & Schaffer, 1978), the rational model of categorization (Anderson, 1991), and a hierarchical Dirichlet process model (Griffiths et al., 2007).