Exact Learning of Preference Structure: Single-peaked Preferences and Beyond

We consider the setting where the members of a society (voters) have preferences over candidates, and the candidates can be ordered on an axis so that the voters' preferences are single-peaked on this axis. We ask whether this axis can be identified by sampling the voters' preferences. For several natural distributions, we obtain tight bounds on the number of samples required and show that, surprisingly, the bounds are independent of the number of candidates. We extend our results to the case where voters' preferences are sampled from two different axes over the same candidate set (one of which may be known). We also consider two alternative models of learning: (1) sampling pairwise comparisons rather than entire votes, and (2) learning from equivalence queries.