Title: Graphical linear models for paired comparison data
Authors: Ori Davidov - University of Haifa (Israel) [presenting]
Abstract: Graphical linear models for paired comparison data will be discussed from both the frequentist and Bayesian perspectives. Estimation methods and large sample theory are described. Specifically, methods for quantifying the uncertainty in ranking procedures are emphasized. The methodology is illustrated using simulated data and applied to two data sets: a network metaanalysis example and to the ranking of teams in the National Basketball Association (NBA).