Inequalities for random utility models, with applications to ranking and subset choice data (Q5937466): Difference between revisions

The aim of this paper is to obtain inequalities that may be used in determining whether ranking or subset choice data are consistent with an independent random utility (IRU) model. The main results are essentially probability inequalities involving order statistics from a set of independent, nonidentically distributed random variables, with applications beyond the context of random utility models. Section 2 presents the definitions necessary to the study of ranking and choice probabilities for the random utility models. In Section 3 there are derived some inequalities useful to check if a set of ranking or subset choice probabilities is incompatible with the class of independent random utility threshold models. The main technique used for the inequalities is association, applicable also to other types of choice probabilities or probabilities of other orderings of random variables. Some of the inequalities are proven to the sharp. Section 4 is devoted to applications of the obtained inequalities to real data sets, in order to assess whether ranking or subset choice data are compatible or not with an IRU model.