To cite the hyper2 package in publications, please use Hankin (2017).
The modern
pentathlon is an
Olympic sport that comprises five different events: fencing, swimming,
horse riding, and running. This short document discusses a dataset
taken from the 2004 Summer
Olympics,
analysing the order statistic for the top seven competitors:
(I have removed diacritics from some of the names).
pentathlon_table <- ordertable(as.matrix(read.table("pentathlon.txt")))
pentathlon_table
## An ordertable:
## shooting fencing swimming riding running
## Moiseev 5 1 1 6 5
## Zadneprovskis 6 2 5 5 1
## Capalini 4 6 2 3 4
## Cerkovskis 3 3 7 7 2
## Meliakh 1 7 4 1 6
## Michalik 2 4 6 2 7
## Walther 7 5 3 4 3
Thus the top line shows that Moiseev came fifth in shooting, first in fencing and swimming, sixth in riding, and fifth in running. The first column shows that the shooting was won by Meliakh (who came first), runner up was Machalik, and so on.
We can present the same information in another way:
ordertable_to_ranktable(pentathlon_table)
## A ranktable:
## c1 c2 c3 c4 c5 c6 c7
## shooting Meliakh Michalik Cerkovskis Capalini Moiseev Zadneprovskis Walther
## fencing Moiseev Zadneprovskis Cerkovskis Michalik Walther Capalini Meliakh
## swimming Moiseev Capalini Walther Meliakh Zadneprovskis Michalik Cerkovskis
## riding Meliakh Michalik Capalini Walther Zadneprovskis Moiseev Cerkovskis
## running Zadneprovskis Cerkovskis Walther Capalini Moiseev Meliakh Michalik
pentathlon <- suppfun(pentathlon_table)
pentathlon_maxp <- maxp(pentathlon)
pentathlon_maxp
## Capalini Cerkovskis Meliakh Michalik Moiseev Walther Zadneprovskis
## 0.186955 0.087915 0.132319 0.114944 0.179307 0.129999 0.168560
pie(pentathlon_maxp)
And we can test the null that the competitors all have the same Plackett-Luce strengths:
equalp.test(pentathlon)
##
## Constrained support maximization
##
## data: pentathlon
## null hypothesis: Capalini = Cerkovskis = Meliakh = Michalik = Moiseev = Walther = Zadneprovskis
## null estimate:
## Capalini Cerkovskis Meliakh Michalik Moiseev Walther Zadneprovskis
## 0.14286 0.14286 0.14286 0.14286 0.14286 0.14286 0.14286
## (argmax, constrained optimization)
## Support for null: -42.626 + K
##
## alternative hypothesis: sum p_i=1
## alternative estimate:
## Capalini Cerkovskis Meliakh Michalik Moiseev Walther Zadneprovskis
## 0.186955 0.087915 0.132319 0.114944 0.179307 0.129999 0.168560
## (argmax, free optimization)
## Support for alternative: -41.868 + K
##
## degrees of freedom: 6
## support difference = 0.75767
## p-value: 0.95847
Following lines create pentathlon.rda, residing in the data/ directory of the package.
save(pentathlon_table,pentathlon,pentathlon_maxp,file="pentathlon.rda")
hyper2 Package: Likelihood Functions for Generalized Bradley-Terry Models.” The R Journal 9 (2): 429–39.