Michele LA ROCCA | Projects

The research aims to implement and study novel resampling techniques for spatio-temporal models, considering both spatial and temporal heterogeneity. The novel procedures should scale well for increasing spatial and temporal dimensions. All the developed algorithms will be implemented in open-source languages such as R or Python.

The research project aims to investigate the effect of ESG scores on stock returns and risk measures. That also constitutes the starting point for portfolio management and performance.

Propose and discuss resampling techniques for model selection and inference that are suitable to deal with massive data

Sviluppo di metodi di Statistical Learning basati su coefficiente di penalità (Matrix Completion) per dati origine-destinazione, per previsione e imputazione di dati mancanti. Applicazione su matrici Input-Output e su dati di flussi di traffico provenienti da telefonia mobile.Sviluppo di metodi di Screening e Variable Selection basati su approccio marginale, quali la Sure Independence Screening,