A0450
Title: Filters based on statistical data depths for robust multivariate inference
Authors: Claudio Agostinelli - University of Trento (Italy) [presenting]
Giovanni Saraceno - University of Trento (Italy)
Abstract: In the classical contamination models, such as the Huber-Tukey contamination model (Case-wise Contamination), observations are considered as the units to be identified as outliers or not. This model is very useful when the number of considered variables is moderately small. It has been shown that the limits of this approach for a larger number of variables and introduced the Independent contamination model (Cell-wise Contamination) where the cells are the units to be identified as outliers or not. One approach to deal, at the same time, with both types of contamination is to filter out the contaminated cells from the data set and then apply a robust procedure able to handle case-wise outliers and missing values. Here we develop a general framework to build filters in any dimension based on statistical data depth functions. We show that previous approaches to construct filters are special cases. We discuss the main theoretical properties of our method and illustrate its performance by Monte Carlo simulation and examples.
