Learning preferences and other remarkable ordinal patterns with finite mixtures of ranking models
:
Venerdì 16 febbraio 2024 ore 12.30, sala 34
Learning preferences and other remarkable ordinal patterns with finite mixtures of ranking models
Prof.ssa Cristina Mollica
Abstract
Ranking data are ubiquitous in various research and applied areas, specifically where the comparison among multiple items represents the main outcome of interest and is recorded in the form of a relative order of the alternatives. Historically, the Mallows model and Plackett-Luce distribution occupy a central role in parametric modelling of ranking data to learn preferences of a population of judges. We present recent methodological and computational advances of the above models to handle partially observed rankings and explore sample heterogeneity through the finite mixture framework. The usefulness of the proposals is illustrated with applications to real data examples from different research contexts.
Per info e link: g.grani@uniroma1.it
Relatore:
Cristina Mollica
Data:
16/02/2024 - 12:30
Luogo:
[Sala 34, piano IV del Dipartimento di Scienze Statistiche, edificio CU002 in Città Universitaria, Piazzale Aldo Moro 5, Roma.]
Allegati:
