Title: Permutation tests for C-sample problems: A multivariate scenario
Authors: Elena Barzizza - Università degli studi di Padova (Italy) [presenting]
Rosa Arboretti - University of Padova (Italy)
Nicolo Biasetton - Università degli Studi di Padova (Italy)
Marta Disegna - University of Padova (Italy)
Luca Pegoraro - University of Padova (Italy)
Luigi Salmaso - University of Padova (Italy)
Abstract: The comparison between multivariate populations, which are characterized by a large number of outcome variables $V$, can be very challenging and particularly interesting in case we have $c>2$ samples (where $c$ is the number of samples considered). In this context, it may be worth considering the application of the Nonparametric Combination (NPC) methodology. The NPC methodology is well known to be flexible and quite powerful in situations characterised by multivariate data and numerous samples. For this reason, we introduce and compare a couple of NPC-based tests to deal with a specific real-world problem, in which we need to order $c>2$ products according to $V>1$ performance measures (i.e. c multivariate samples need to be compared). The application of these testing procedures allows us to demonstrate the flexibility of the Nonparametric Combination methodology and how it can be adopted to address a really common industrial problem.