sonia.petrone@unibocconi.it
Bio
I am a Professor of Statistics at Bocconi University, Milan. I graduated in Discipline Economiche e Sociali at Bocconi, and obtained my PhD in Statistics in 1989. I held positions at the University of Pavia (1991-1998) and University of Insubria (1998-2001), before joining Bocconi in 2001. Throughout my career, I conducted extensive research visits at various universities across the USA, Latin America, Europe, India, and Russia.
About
At Bocconi, I played an active role in the development of the PhD in Statistics, as Vice-Director and Curricula Responsible (2004-2011) and Director (2011-2018). I am the Director of the Bocconi Summer School in Advanced Statistics and Probability since 2017.
I was the Editor of Statistical Science (2020-2022), and a co-Editor of Bayesian Analysis (2010-2014).
I was the President of the International Society for Bayesian Analysis (ISBA) in 2014 (President-Elect 2013, Past President 2015). My professional service includes having been an elected member of the ISBA Board of Directors (2002-2004 and 2008-2010) and of the Institute of Mathematical Statistics (IMS) Council (2011-2014 and ex-officio 2020-2022). I was member (2020-204, and currently ex-officio) and Chair of the ERC of the Bernoulli Society (2023-2024)
I am an elected Fellow of ISBA and an IMS Fellow, and an elected member of ISI. I have been awarded an IMS Medallion Lecture in 2018 and an ISBA Foundational Lecture in 2016.
I am an elected Fellow of the European Laboratory for Intelligent Systems (ELLIS), and a Fellow of the Bocconi Institute of Data Science (BIDSA).
Research Interests
I am a Bayesian statistician and my research interests encompass foundational theory as well as methods and applications of Bayesian Statistics.
Key areas in my work on foundations, in the legacy of Bruno de Finetti, include the role of prediction, exchangeability and partial exchangeability, predictive characterizations under and beyond exchangeability, stochastic processes with reinforcement. More broadly, my research advocates a predictive approach to (Bayesian) statistics, which has substantial implications for a wide range of problems in statistics and machine learning.
I am interested in the broad area of Bayesian nonparametrics. My research work goes from developing prior laws for random functions, for instance using random Berstein polynomials and mixture models, to nonparametric methods for hierarchical structures, clustering, high-dimensional regression, spatial and spatio-temporal data, and latent variable models. I am also interested in Bayesian and frequentist properties of Bayesian (nonparametric) procedures.
My research interests extend to state space models, and applications in neurosciences, empirical Bayes and Bayes empirical Bayes methods, and statistical machine learning.