Person
ZUCCOLOTTO Paola

ZUCCOLOTTO Paola

Docenti di ruolo di Ia fascia
Department of Economics and Management
Course Catalogue:
https://unibs.coursecatalogue.cineca.it/docente-ma...
SCIENTIFIC PROFILE She carries out scientific research activity in the field of Statistical Sciences, both with a methodological and applied approach. Main topics: - Methodological and empirical aspects of data mining and data analysis. From a methodological point of view, she is interested to computer intensive algorithmic techniques such as neural networks and learning ensembles, with special attention the problem of variable selection, for which she has developed some innovative proposals. In addition, she has proposed some novel techniques for missing data treatment in Principal Component Analysis and for modelling rating data coming from surveys about human perceptions. From an empirical point of view, she has been concerned with several case studies in the framework of perceptions measurement, quality evaluation, prediction and risk factors identification within different contexts (healthcare, genetics, job, marketing, sensory analysis, sport). - Time series analysis, with special attention to financial data. In this framework she carried out researches on vector time series, models for the volatility of financial data, models for changes in regime and she has proposed an innovative model for ultra-high-frequency data. At the moment she is attempting to extend data mining techniques to econometrics, with a special link to the theory of copula functions from the perspective of investment risk management. She published more than 50 peer-reviewed papers on national and international journals and books, one book and more than 30 working papers and conference proceedings. She participated to more than 50 international conferences, as invited speaker, contributed speaker or invited discussant. She is and has been member of several research projects, granted by national (PRIN) and international (European Union Seventh Framework Programme) funds. PROFESSIONAL PROFILE Full Professor of Statistics at the University of Brescia. She is the scientific director of the Big & Open Data Innovation Laboratory (BODaI-Lab), where she coordinates, together with Marica Manisera, the international project Big Data Analytics in Sports (BDsports). She carries out scientific research activity in the field of Statistical Sciences, both with a methodological and applied approach. She regularly acts as scientific reviewer for the most prestigious world journals in the field of Statistics.She is a member of the Editorial Advisory Board of the Journal of Sports Sciences and guest co-editor of special issues of international journals on Statistics in Sports. She has been delegated by the International Statistical Institute the task of constituting a new Special Interest Group on Sports Statistics. She teaches undergraduate and graduate courses in the field of Statistics and is responsible for the scientific area dedicated to Sport Analytics at the PhD "Analytics for Economics and Management" of the University of Brescia. She also teaches courses and seminars on Sports Analytics in University Masters on Sports Engineering and specialized training projects devoted to people operating in the sports world. She supervises students’ internships, final reports and master's theses on the subject of Statistics and works in collaboration with high-school teachers, creating experimental educational projects to bring students closer to quantitative subjects through Sport Analytics.

Fields (3)

Free text keywords (2)

DATA ANALYSIS
STATISTICS
No Results Found

Research fields (4)

Sport Analytics: development of statistical tools in the fields of statistical modelling, multivariate data analysis, data mining, algorithmic modelling and machine learning. Main topics: − Basic statistics and more complex analytics of a match or a competition − Performance analysis (of teams, players, individual athletes) − Identification of success factors and optimal game strategies − Forecasting − Personality traits measurement (mental toughness, coping strategies, …) − Market analysis for sport marketing − Financial assessment of sports clubs and sports related projects
Study of data mining techniques, with specific regard to: - Algorithmic techniques for regression and classification, with special interest in Neural Networks, classification and regression trees (CART), tree-based learning ensembles such as Random Forests (Breiman, 2001) and Gradient Boosting Machine (Friedman , 2002). - Techniques for variable selection, with particular reference to variable importance measures. SCIENTIFIC CONTRIBUTIONS: development of innovative techniques in the context of variable importance measurement and correction procedures for the distortion of the Gini variable importance measure.
Study of issues related to time series, with specific regard to: - Stochastic models for the treatment of heteroskedasticity, a particularly relevant problem in the analysis of financial time series, with special attention to the ARCH model class proposed by Engle (1982). - Stochastic models with multiple regimes, for the treatment of economic and financial time series characterized by asymmetric alternations of periods with different characteristics (expansion / recession, quiet / turbulence, ...). Special attention has been paid to the Markov Switching models. - Multivariate time series analysis. Among the several issues related to the analysis of multivariate time series, the VARMA linear modeling and especially the problem of the construction and estimation of canonical forms was studied in depth. - Analysis of financial data collected at very high frequency, with particular reference to the ACD models of Engle and Russell (1998). - Analysis of financial data through the copula function approach. - Clustering of financial time series. SCIENTIFIC CONTRIBUTIONS: as part of the ACD models, extensions and in-depth analysis of the model structure, distribution hypotheses, and methods for parameter estimation; innovative financial time series clustering techniques, based on estimated tail dependency coefficients with copula functions.
Study of multivariate data analysis techniques, with specific regard to: - Dimensionality reduction and cluster analysis techniques. - Statistical models for latent variables. - Statistical models for the analysis of ordinal data, with particular attention to the CUB class. - Techniques for the Analysis of Symbolic Data. SCIENTIFIC CONTRIBUTIONS: development of innovative techniques for modeling data and dealing with missing values, within the CUB models; treatment of missing values ​​through the techniques of Symbolic Data Analysis.
No Results Found

Other research activities

Advanced Clustering for Human perceptions in Integrated Views of wEllbeing 
Bando Cariplo
Project
Scientific Manager
2027
36 months
No Results Found

Publications (155)

Office Hours

Video

Useful Resources