1. Statistical Model Selection and Comparison
His research addresses the development and assessment of methods for statistical model selection and comparison, with particular emphasis on robustness under model misspecification. Topics include information criteria, asymptotic and finite-sample properties, consistent model selection, and predictive performance. Part of this research builds on work on the Misspecification-Resistant Information Criterion (MRIC) and its extensions to multivariate models and time series.
2. Model Misspecification, Covariance Modelling and Statistical Inference
A current line of research investigates the consequences of model misspecification for statistical inference, with particular emphasis on covariance structures. The work studies which quantities are effectively recovered when the assumed model differs from the data-generating process and how alternative assumptions on covariance, error distributions, and model structure affect estimation, inference, and model comparison.
3. Statistical Methodology for Cosmological Applications
At the National Institute for Nuclear Physics (INFN), Ferrara Division, his research focuses on the development and assessment of statistical methodology for cosmological inference. Current topics include model comparison, robustness under misspecification, covariance estimation, uncertainty propagation, and the inferential properties of statistical procedures used in cosmological data analysis.
4. Time Series Analysis and Forecasting
His research on time series concerns multivariate modelling, forecasting, model selection, and predictive-error assessment. Particular attention is given to finite-sample properties, vector and dynamic models, and the robustness of model-selection and forecasting procedures.
5. Econometrics and Applied Statistical Modelling
Part of his research concerns the application of statistical and econometric methods to empirical problems in economics, finance, and other applied fields. Topics include dynamic and panel-data models, causal inference, longitudinal data, microdata, spatial econometrics, and quantitative analysis of economic and financial phenomena. Previous applications also include high-frequency sports data and behavioural data.
