Modelling rankings in R: the PlackettLuce package

10/29/2018 ∙ by Heather L. Turner, et al. ∙ Heather Turner 0

This paper presents the R package PlackettLuce, which implements a generalization of the Plackett-Luce model for rankings data. The generalization accommodates both ties (of any order) and partial rankings (rankings of only some items). By default, the implementation adds a set of pseudo-comparisons with a hypothetical item, ensuring that the network of wins and losses is always strongly connected, i.e. all items are connected to every other item by both a path of wins and a path of losses. This means that the worth of each item can always be estimated by maximum likelihood, with finite standard error. It also has a regularization effect, shrinking the estimated parameters towards equal item worth. In addition to standard methods for model summary, PlackettLuce provides a method to compute quasi standard errors for the item parameters, so that comparison intervals can be derived even when a reference item is set. Finally the package provides a method for model-based partitioning using ranking-specific covariates, enabling the identification of subgroups where items have been ranked differently. These features are demonstrated through application to classic and novel data sets.



There are no comments yet.


page 1

page 2

page 3

page 4

This week in AI

Get the week's most popular data science and artificial intelligence research sent straight to your inbox every Saturday.

1 Introduction

2 Methods

3 Plackett-Luce Trees

4 Discussion

5 Appendix


  • Allison and Christakis (1994) Allison PD, Christakis NA (1994) Logit Models for Sets of Ranked Items. Sociological Methodology 24:199, DOI 10.2307/270983, URL
  • Baker (1994) Baker SG (1994) The multinomial-Poisson transformation. Journal of the Royal Statistical Society Series D (The Statistician) 43(4):495–504, DOI 10.2307/2348134
  • Coppola et al (2014) Coppola A, Stewart B, Okazaki N (2014) lbfgs: Limited-memory BFGS Optimization. URL, R package version 1.2.1
  • Davidson (1970) Davidson RR (1970) On extending the bradley-terry model to accommodate ties in paired comparison experiments. Journal of the American Statistical Association 65(329):317–328, DOI 10.1080/01621459.1970.10481082
  • Firth (2003) Firth D (2003) Overcoming the reference category problem in the presentation of statistical models. Sociological Methodology 33(1):1–18, DOI 10.1111/j.0081-1750.2003.t01-1-00125.x
  • Firth and De Menezes (2004) Firth D, De Menezes RX (2004) Quasi-variances. Biometrika 91(1):65–80, DOI 10.1093/biomet/91.1.65
  • Hankin (2018) Hankin RKS (2018) hyper2: The Hyperdirichlet Distribution, Mark 2. URL, R package version 1.0-4
  • Hunter (2004) Hunter DR (2004) MM algorithms for generalized Bradley-Terry models. Annals of Statistics 32(1):384–406, DOI 10.1214/aos/1079120141, arXiv:1210.1016v1
  • Lee and Yu (2015) Lee PH, Yu PLH (2015) pmr: Probability Models for Ranking Data. URL, R package version 1.2.5
  • Luce (1959) Luce RD (1959) Individual Choice Behavior: A Theoretical Analysis. New York: Wiley
  • Luce (1977) Luce RD (1977) The choice axiom after twenty years. Journal of Mathematical Psychology 15(3):215–233, DOI 10.1016/0022-2496(77)90032-3
  • Mattei and Walsh (2013)

    Mattei N, Walsh T (2013) PrefLib: A Library of Preference Data. In: Proceedings of the 3rd International Conference on Algorithmic Decision Theory (ADT 2013), Springer, Lecture Notes in Artificial Intelligence, URL
  • Maystre (2018) Maystre L (2018) choix: Inference algorithms for models based on Luce’s choice axiom. URL, python package version 0.3.2
  • Mollica and Tardella (2018) Mollica C, Tardella L (2018) PLMIX: An R package for modeling and clustering partially ranked data, URL, 1612.08141
  • Plackett (1975) Plackett RL (1975) The Analysis of Permutations. Appl Statist 24(2):193–202, DOI 10.2307/2346567
  • Seng and Yilin (2018) Seng TC, Yilin N (2018) ROlogit: Fit Rank-Ordered Logit (RO-Logit) Model. URL, R package version 0.1.1
  • Strobl et al (2009) Strobl C, Malley J, Tutz G (2009) An introduction to recursive partitioning: rationale, application, and characteristics of classification and regression trees, bagging, and random forests. Psychological methods 14(4):323–48, DOI 10.1037/a0016973, URL
  • Strobl et al (2011) Strobl C, Wickelmaier F, Zeileis A (2011) Accounting for Individual Differences in Bradley-Terry Models by Means of Recursive Partitioning. Journal of Educational and Behavioral Statistics 36(2):135—-153, DOI 10.3102/1076998609359791, URL
  • Turner and Firth (2018) Turner H, Firth D (2018) Generalized nonlinear models in R: An overview of the gnm package. URL, R package version 1.1-0
  • Turner et al (2018) Turner H, Kosmidis I, Firth D (2018) PlackettLuce: Plackett-Luce Models for Rankings. URL, R package version 0.2-3
  • Van Etten et al (2016) Van Etten J, Beza E, Calderer L, van Duijvendijk K, Fadda C, Fantahun B, Kidane YG, van de Gevel J, Gupta A, Mengistu DK, Kiambi D, Mathur PN, Mercado L, Mittra S, Mollel MJ, Rosas JC, Steinke J, Suchini JG, Zimmerer KS (2016) First experiences with a novel farmer citizen science approach: crowdsourcing participatory variety selection through on-farm triadic comparisons of technologies (tricot). Experimental Agriculture pp 1–22, DOI 10.1017/S0014479716000739, URL{_}S0014479716000739
  • Wang and Erosheva (2015) Wang YS, Erosheva EA (2015) mixedMem: Tools for Discrete Multivariate Mixed Membership Models. URL, R package version 1.1.0