Aggregation of Rankings: a Brief Review of Distance-Based Rules

Abstract

Some researchers have addressed the problem of aggregating individual preferences or rankings by seeking a ranking that is closest to the individual rankings. Their methods differ according to the notion of distance that they use. The best known method of this sort is due to Kemeny. The first part of this paper offers a brief survey of some of these methods. Another way of approaching the aggregation of rankings is as a problem of optimal statistical inference, in which an expected loss is minimised. This approach requires a loss function, a concept closely related the notion of distance between rankings. The second part of this paper examines two classes of parametric functions and proposes one class for the optimal statistical inference problem.