A0237
Title: Bootstrap inference for signal reconstruction from multiple ranked lists
Authors: Michele La Rocca - University of Salerno (Italy) [presenting]
Bastian Pfeifer - Medical University of Graz (Austria)
Michael Georg Schimek - Medical University of Graz (Austria)
Abstract: Statistical ranking procedures are widely used to rate objects' relative quality or relevance across multiple assessments. Beyond rank aggregation, estimating the usually unobservable latent signals that inform a consensus ranking is interesting. Under the only assumption of independent assessments, we have introduced an indirect inference approach via convex optimisation, which is computationally efficient even when the number of assessors is much lower than the number of objects to rank. Notably, the novel signal estimator can be written as a weighted estimator, which opens the possibility of using weighted bootstrap schemes to implement efficient resampling procedures. Those procedures are key in gaining inference on the unknown signals (by computing standard errors or confidence intervals) or testing signal differences between groups. Within this framework, we compare alternative weighted bootstrap schemes (namely, Poisson, Multinomial and Dirichlet) for their different computational burden and ability to approximate the unknown sampling distribution of the signal estimators accurately. The bootstrap procedures will be evaluated, by Monte Carlo simulation, on different scenarios with increasing problem complexity, including several combinations of the number of assessors and the number of objects to rank.
