| Popis: |
We model stochastic choices with categorization, resulting from the preliminary step of grouping alternatives in homogenous disjoint classes. The agent randomly chooses one class among those available, then randomly picks an item within the selected class. We give a formal definition of a choice generated by this procedure, and provide a characterization. The characterizing properties allow an external observer to elicit that categorization is applied. In a more general interpretation, the model allows to describe the observed choice as the composition of independent subchoices. This composition preserves rationalizability by random utility maximization. A generalization of the model subsumes Luce model and Nested Logit. |
