A0986
Title: M-quantile regression shrinkage and selection via the Lasso and Elastic Net
Authors: Francesco Pantalone - University of Southampton (United Kingdom) [presenting]
Maria Giovanna Ranalli - University of Perugia (Italy)
Nicola Salvati - University of Pisa (Italy)
Lea Petrella - Sapienza University of Rome (Italy)
Abstract: An M-quantile regression model with Lasso and Elastic Net penalizations is presented. This new methodology allows (i) to identify the best predictors via model selection, (ii) to investigate the relationship between response and covariates at different M-quantiles of the conditional response distribution, and (iii) to be robust to the presence of outliers. Finally, heterogeneity in the data can be accounted for via B-spline. A real application of the effect of traffic on air quality in the city of Perugia (Italy) is presented.
