Open Conference Systems, 50th Scientific meeting of the Italian Statistical Society

Font Size: 
Clustering of preference rankings: a non-parametric soft-clustering approach
Antonio D'Ambrosio, Willem J Heiser

Last modified: 2018-03-23


Typically, rank data consist of a set of individuals, or judges, who have
ordered a set of items or objects according to their overall preference or some prespecified criterion. When each judge has expressed his or her preferences according to his own best judgment, such data are characterized by systematic individual
differences. In the literature several approaches have been proposed in order to decompose heterogeneous populations into a defined number of homogeneous groups.
Often, these approaches work by assuming that the ranking process is governed by some distance-based models.
We use the flexible class of methods proposed by Ben-Israel and Iyigun, which consists in a probabilistic-distance clustering approach, and define the disparity between a ranking and the center of a cluster as the Kemeny distance. This class of methods allows for probabilistic allocation of cases to classes, being a form of fuzzy clustering, rather than hard clustering, where the probability is unequivocally related to the chosen distance measure.

Full Text: PDF