Predictions with Uncertain Categorization: A Rational Model

A key function of categories is to help predictions about unobserved features of objects. At the same time, humans often find themselves in situations where the categories of the objects they perceive are uncertain. How do people make predictions about unobserved features in such situations? We propose a rational model that solves this problem. Our model complements existing models in that it is applicable in settings where the conditional independence assumption does not hold (features are correlated within categories) and where the features are continuous as opposed to discrete. The qualitative predictions of our model are borne out in two experiments.