Track Policies

Session Code

SPEC: Specialized
SSPEC: Super specialized
SOLI: Solicited
CONT: Contributed
BES: The BES Day - Special event in collaboration with ISTAT
Mx.y: Morning, day, number
Ax.y: Afternoon, day, number

Unchecked Open Submissions Unchecked Peer Reviewed

Plenary Session

M1: Three-Mode Analysis Problems and Prospects
Speaker: Pieter Kroonenberg, University of Leiden
Chair: Rosaria Lombardo

M2: Factorial methods for temporal data
Speaker: Renato Coppi, Sapienza  University of  Rome
Chair: Giuseppe Bove

M3: The role of fixed-effects approaches in latent variable modeling
Speaker: Anders Skrondal,  Norwegian Institute of Public Health
Chair: Carla Rampichini

A1: Advances in estimation of large hierarchical spatio-temporal models
Speaker: Havard Rue, Norwegian University of Science and Technology
Chair: Daniela Cocchi

A3-BES: Measuring the progress of societies: a key tool to improve policy making and societal well-being
Speakers: Marleen De Smedt, Eurostat;   Linda Laura Sabbadini, ISTAT;   Maria Teresa Salvemini, CNEL
Chair: Cristina Freguja
Measuring progress of our societies - A joint undertaking in the European Statistical System (ESS)
Measuring weel-being in Italy: the new challanges of BES

The economic approach to the BES

Directors
  • Silvia Golia, University of Brescia
  • Marica Manisera, University of Brescia
  • Marika Vezzoli, University of Brescia
  • Paola Zuccolotto, University of Brescia
Checked Open Submissions Checked Peer Reviewed

Super Specialized Session

SSPEC-A2.1: Factor models for multivariate time series analysis
Speaker: Daniel Peña, Universidad Carlos III de Madrid;  Marco Lippi,  University of Roma La Sapienza
Chair: Angela Montanari
Dimension reduction in the time domain: Dynamic Principal Components
Dynamic Factor Models with Infinite-Dimensional Factor Space: One-Sided Representations

SSPEC-A2.2: Uncovering the latent structure in sensory data
Speaker: El Mostafa Qannari, LUNAM University
Chair: Pietro Amenta

BES-A3.1: The common path to face the challenges of equitable and sustainable well-being
Speakers: Giorgio Sirilli, CNR;  Mauro Agnoletti, University of Florence;  Saverio Gazzelloni, Istat
Chair: Innocenzo Cipolletta
Research and Innovation in the measurement of well-being
The value of the landscape in measuring well-being
Culture and well-being: the develpoment of a conceptual framework

Directors
  • Silvia Golia, University of Brescia
  • Marika Vezzoli, University of Brescia
  • Paola Zuccolotto, University of Brescia
Checked Open Submissions Checked Peer Reviewed

BES-M3.1 - The BES and the challenges of constructing composite indicators dealing with equity and sustainability

The session faces the technical challenges of constructing composite indicators dealing with two important dimensions of the BES project, equity and sustainability. The cases of equity in work through job quality and of well-being polarization in Italian society are considered. Finally a Comparison of Unbalance Adjustment Methods is proposed.

Directors
  • Maurizio Carpita, University of Brescia
  • Silvia Golia, University of Brescia
  • Filomena Maggino, Università degli Studi di Firenze Dipartimento di Statistica, Informatica, Applicazioni "G. Parenti" (DiSIA)
  • Elena Tosetto, Statistics Directorate - OECD
Checked Open Submissions Checked Peer Reviewed

BES-M3.2 - The integration of equity, vulnerability and sustainability: the main challenges of well-being

The session focuses on theoretical and methodological problems that arise in measuring equitable and sustainable well-being in Italy and, in this perspective the main challenge is to integrate equity, vulnerability and sustainability. In the first report on equitable and sustainable well being, indicators of inequality were proposed jointly with well-being indicators for a number of structural variables. Inequalities were considered also for nontraditional economic aspects to better allow the identification of excluded groups and of lack of opportunities cases.  The problem of how to integrate sustainability into BES remains open.

Directors
  • Maurizio Carpita, University of Brescia
  • Saverio Gazzelloni, Istituto nazionale di Statistica
  • Silvia Golia, University of Brescia
Checked Open Submissions Checked Peer Reviewed

SPEC-M1.1 - Space and space-time models A: methods and environmental applications

1) A spatiotemporal latent factor model estimated by the EM algorithm is developed, estimation problems solved, and used to cluster water quality of lakes around the world.
2) A Bayesian spatio-temporal classification method based on finite mixture models is developed and used to classify lagoons according to multiple biotical indicators.
3) Spatiotemporal covariances for large datasests

Directors
  • Alessandro Fasso, University of Bergamo
  • Silvia Golia, University of Brescia
  • Marco Minozzo, University of Verona
Checked Open Submissions Checked Peer Reviewed

SPEC-M1.2 - Latent variable models for marketing research

Because of the empirical nature of research in marketing, the analyst often faces measurement or, more generally, quantitative problems. This means that the so-called "quantitative marketing" is strongly related to the use of statistical and econometric techniques. The session focuses on some advanced latent variables methodologies that have been recently introduced in the marketing literature and are particularly qualified to provide suitable solutions to several problems regarding the definition and the development of marketing strategies.

Directors
  • Francesca Bassi, University of Padova
  • bruno scarpa, Dipartimento di Scienze Statistiche - Università di Padova
  • Marika Vezzoli, University of Brescia
Checked Open Submissions Checked Peer Reviewed

SPEC-M2.1 - Dynamic Models with Latent Variables

Estimation and inference of latent variable models for time series have a longstanding tradition in both theoretical and empirical researches and cover a wide range of academic and operational fields. In many situations an inference procedure has to deal with the unobserved latent variables, which typically involves either using proxies for the latent variables or integrating out the latent variables in the likelihood function. This is the typically the case in economic and financial time series given the latent nature of the major indicators. The session aim at investigate and present same recent development into the field.

Directors
  • Alessandra Amendola, Universita di Salerno
  • Marcella Corduas, Università di Napoli Federico II
  • Paola Zuccolotto, University of Brescia
Checked Open Submissions Checked Peer Reviewed

SPEC-M2.2 - Issues in estimating complex latent trait and latent class models

The session is dedicated to new tools for estimating complex models. In particular the papers deal with a weighted pairwise likelihood estimator for a class of latent variable models, a Bayesian estimation of IRT models with power priors and a multilevel mixture factor models in an IRT parameterization.

Directors
  • Silvia Cagnone, Department of Statistics University of Bologna
  • Marica Manisera, University of Brescia
  • Stefania Mignani, Alma Mater Studiorum - Università di Bologna
  • Fulvia Pennoni, Università degli Studi di Milano Bicocca
Checked Open Submissions Checked Peer Reviewed

SPEC-M2.3 - Sensorial data analysis

This session aims at drawing an overview of the use of advanced statistical methods based on Latent Variables in food industry, evaluating practical benefits given by these new techniques. Main attention is given on multivariate statistical analyses for studying: expert panel data or consumer data, the relationships between instrumental and sensorial measurements and between consumer preferences and sensorial data.

Directors
  • Rosaria Lombardo, Second University of Naples
  • Walter Maffenini, Dipartimento di Statistica e Metodi Quantitativi
  • Marica Manisera, University of Brescia
Checked Open Submissions Checked Peer Reviewed

SPEC-M2.5 - Essays in Anchoring Vignettes

This session provides an overview of the anchoring vignette approach. Vignettes are innovative types of questionnaire, introduced less than ten years ago for analyzing quantities that are not directly observable (i.e. self-evaluations), controlling for the heterogeneity of response styles among respondents. The econometric specification is called “hopit” model; it consists of two components, jointly modelling selfassessments and vignette evaluations through standard ordered Probit models.

Directors
  • Francesca Bassi, University of Padova
  • Omar Paccagnella, Department of Statistical Sciences, University of Padua
  • Silvia Salini, Univeristy of Milan
  • Marika Vezzoli, University of Brescia
Checked Open Submissions Checked Peer Reviewed

SPEC-M3.1 - Latent variables in demographic analysis

Many demographic studies need to take into account individual heterogeneity by mean of latent factor models. Some applications (ranging from mortality to family formation) are proposed in the session.

Directors
  • Marica Manisera, University of Brescia
  • Stefano Mazzuco, University of Padova
  • Daniele Vignoli, Dipartimento di Statistica "G. Parenti", Università di Firenze
Checked Open Submissions Checked Peer Reviewed

SPEC-M3.2 - Special symbolic data analysis

With the advent of computers, large, very large datasets have become routine. What is not so routine is how to analyse these data and/or how to glean useful information from within their massive confines. One approach is to summarize large data sets in such a way that the resulting summary dataset is of a manageable size. One consequence of this is that the data may no longer be formatted as single values such as is the case for classical data, but may be represented by lists, intervals, distributions and the like. These summarized data are examples of symbolic data. This session looks at the concept of special symbolic data, and then attempts to new methodologies to analyse such data.

Directors
  • Silvia Golia, University of Brescia
  • Maria Gabriella Grassia, Dipartimento di Scienze Sociali - Università degli Studi Federico II di Napoli
Checked Open Submissions Checked Peer Reviewed

SPEC-M3.3 - Structural Equation Models, Covariance Structure Analysis and Partial Least Squares

The drawbacks of the SEM have extensively discussed, and in many cases the PLS-PM approach is preferred. PLS-PM suffers no indeterminacy problems, it requires no assumption of multinormality in the data and it is efficiently implemented in many software packages. Although it is extensively applied to formative-reflective models and also to purely reflective models formally represents latent variables in a formative way. CSA can be proposed as alternative method to overcome these problems.

Directors
  • Maurizio Vichi, DSPSA
  • Giorgio Vittadini, Università di Milano Bicocca
  • Paola Zuccolotto, University of Brescia
Checked Open Submissions Checked Peer Reviewed

SPEC-M3.4 - Multi-way Component Analysis

The need for multi-way analysis arises when data are arranged in three- or higher way arrays. Multi-way techniques, such as the popular Tucker3 and Candecomp/Parafac models, synthesize data arrays by means of latent variables and, thus, can be seen as natural extensions of classical factor analysis. In fact, multi-way models are multi-linear generalizations of the standard bi-linear model of factor analysis (for two-way data) where extra sets of factor loadings are incorporated for the additional ways of the data array.

Directors
  • Giuseppe Bove, University Roma Tre
  • Renato Coppi, Sapienza Università di Roma
  • Paolo Giordani, Sapienza University of Rome
  • Marika Vezzoli, University of Brescia
Checked Open Submissions Checked Peer Reviewed

SPEC-M3.5 - Finite mixture models for categorical variables

In the statistical literature there has been a growing interest in finite mixture modeling as a tool to increase the flexibility of conventional parametric models. Finite mixture models can be seen as a compromise between a simple parametric model and a non-parametric approach. Mor eover, these models allow to account for unobserved heterogeneity due to latent sub-populations, often called latent classes. This session will focus on recent advances categorical or discrete.

Directors
  • Ruggero Bellio, Università di Udine
  • Carla Rampichini, Department of Statistics, Computer Science, Applications 'G. Parenti', Università di Firenze
  • Paola Zuccolotto, University of Brescia
Checked Open Submissions Checked Peer Reviewed

SPEC-A1.1 - Finite mixture models for complex data structure

The session will include new mixture-based methods for latent variable models with nonnormal random effects.

Directors
  • Francesco Lagona, University Roma Tre
  • Marica Manisera, University of Brescia
  • Antonello Maruotti, Libera Università Maria Ss. Assunta University of Southampton
Checked Open Submissions Checked Peer Reviewed

SPEC-A1.2 - The latent variables hidden in the texts

The classic text mining approach, ie the application of data mining techniques to 'unstructured' text documents in order to 'select' information, has evolved in many directions in recent years. The aim is always the extraction of latent information contained in a set of documents. The session would include some of the most significant developments, in particular the Topic Models.

Directors
  • Pier Alda Ferrari, University of Milano
  • Silvia Golia, University of Brescia
Checked Open Submissions Checked Peer Reviewed

SPEC-A1.3 - Graphical Markov representation of latent structures

Graphical Markov models have been widely used as a representation of conditional independence in machine learning and statistics. A key feature of graphical models is that they can be used to represent the independence structure among a set of observed variables, assuming they are part of a larger structure that includes hidden variables. The latter is one of the most active research areas in graphical modelling and this Session aims to present recent advances in this direction.

Directors
  • Luca La Rocca, University of Modena and Reggio Emilia
  • Angela Montanari, Alma Mater Studiorum - University of Bologna
  • alberto roverato, Università di Bologna
  • Paola Zuccolotto, University of Brescia
Checked Open Submissions Checked Peer Reviewed

SPEC-A1.4 - Latent variables in causal inference

The session will focus on the specific challenges for causal inference arising when units are clustered in latent heterogeneous sub-populations. Methodological issues and substantive empirical studies related to the identification and estimation of causal effects in those complicated settings will be discussed.

Directors
  • Fabrizia Mealli, University of Firenze
  • Carla Rampichini, Department of Statistics, Computer Science, Applications 'G. Parenti', Università di Firenze
  • Elena Stanghellini, Dipartimento di Economia, Finanza e Statistica - Università di Perugia
  • Paola Zuccolotto, University of Brescia
Checked Open Submissions Checked Peer Reviewed

SPEC-A2.1 - Latent Variables and Causal Inference in Genetics

The contributions scheduled in this session present some approaches to genetic data analysis by means of statistical techniques in the framework of latent variables modelling, as well as instrumental variable approaches for causal inference based on the use of genes. The topics widely range from algorithmic techniques applied in the context of genome-wide association studies to methods for assessing causality of the relationship between modifiable risk factors (e.g. biomarkers) and disease.

Directors
  • Cosetta Minelli, Respiratory Epidemiology and Public Health, National Heart& Lung Institute, Imperial College, London
  • Mario Pezzotti, University of Verona
  • Paola Zuccolotto, University of Brescia
Checked Open Submissions Checked Peer Reviewed

SPEC-A2.2 - Exploratory data analysis of contingency tables for latent variables

New approaches to the correspondence analysis for studying latent structure of data on nominal and ordinal scale.

Directors
  • Pietro Amenta, University of Sannio (Benevento, Italy)
  • Giuseppe Boari, Università Cattolica del Sacro Cuore
  • Luigi D'Ambra
  • Marika Vezzoli, University of Brescia
Checked Open Submissions Checked Peer Reviewed

SPEC-A2.3 - Statistics on Manifolds

Statistical analysis on manifolds is a relatively new domain at the confluent of several mathematical and application domains. Its goal is to provide a statistical study of geometric objects living in differential manifolds. Focusing on Riemannian manifolds, the three talks of this section will deal with problems on inference, modeling and dimensionality reduction. The statistical analysis will consider applications in directional and axial statistics, morphometrics, medical diagnostics and machine vision.

Directors
  • Marco Alfo', Dipartimento di scienze Statistiche, Sapienza Università di Roma
  • Tonio Di Battista, Università G. d'Annunzio di Chieti-Pescara
  • Luigi Ippoliti, University of Chieti-Pescara
  • Marica Manisera, University of Brescia
Checked Open Submissions Checked Peer Reviewed

SPEC-A2.4 - PLS path modeling on structured data

When applying structural equation modeling methods, such as PLS path modeling, in empirical studies, the assumption that the data have been collected from a single homogeneous population is often unrealistic. Unobserved heterogeneity in PLS estimates on the aggregate data level may result in misleading interpretation. This problem can be solved through different classification methods. Supervisioned and Unsupervisioned Classification, in PLS path modeling, will be discussed.

Directors
  • Luigi Fabbris, Università di Padova
  • Silvia Golia, University of Brescia
  • Maria Gabriella Grassia, Dipartimento di Scienze Sociali - Università degli Studi Federico II di Napoli
  • Natale Lauro, Dipartimento di Scienze Economiche e Statistiche, Università Federico II di Napoli
Checked Open Submissions Checked Peer Reviewed

INVI-M2.1 - Latent Models

Directors
  • Maurizio Carpita, University of Brescia
  • Marica Manisera, University of Brescia
  • Maurizio Vichi, DSPSA
Checked Open Submissions Checked Peer Reviewed

SOLI-M1.1 - Latent models in financial risk management

The session is devoted to the role of latent variables in multivariate statistical models to measure financial risk.

Directors
  • Francesca Bassi, University of Padova
  • Paolo Giudici, University of Pavia
  • Marika Vezzoli, University of Brescia
Checked Open Submissions Checked Peer Reviewed

SOLI-M2.1 - Measures of higher education human capital

The human capital created by universities on graduates is a latent variable. It can be measured in various ways. Measures can simply refer to knowledge growth at the end of study programmes, or it can refer to capabilities implemented on graduates that can be assessed in the labour market, in the productive world and in the civil society at large. Hence, human capital is supposed to be a unique paradigm whose measures could change according to graduates’ life experience and ways of measuring it.

Directors
  • Luigi Fabbris, Università di Padova
  • Silvia Golia, University of Brescia
  • Maria Gabriella Grassia, Dipartimento di Scienze Sociali - Università degli Studi Federico II di Napoli
Checked Open Submissions Checked Peer Reviewed

SOLI-M2.2 - Latent variables in medicine

The role of latent variables and statistical models in medicine is becoming more and more important.This session aims at discussing the most recent contributions to this field.

Directors
  • Pier Alda Ferrari, University of Milano
  • Silvia Golia, University of Brescia
  • John Thompson, University of Leicester
Checked Open Submissions Checked Peer Reviewed

SOLI-M2.3 - Advances in Spatial Econometrics

The session aims at discussing methods and applications concerning the analysis of geographically distributed observations with particular reference to latent factors underlying spatial data.

Directors
  • Luigi Ippoliti, University of Chieti-Pescara
  • Marica Manisera, University of Brescia
  • Paolo Postiglione, University G. d'Annunzio, Chieti-Pescara
Checked Open Submissions Checked Peer Reviewed

SOLI-M2.4 - Sensory Analysis in action

Statisticians usually have a very vague idea of the tools used in sensory analysis. They, however, have definite ideas on statistical instruments. Those involved in sensory analysis know rather well their own instruments, but tend to always use the same statistical tools.
In this session we will try to meet the statistical world and the sensory analysis to improve the knowledge and therefore the quality of the sensory survey’s results.

Directors
  • Eugenio Brentari, University of Brescia
  • Marica Manisera, University of Brescia
Checked Open Submissions Checked Peer Reviewed

SOLI-M3.1 - Scale reliability for ordinal item responses

More and more often, in social sciences studies, data coming from questionnaire surveys, are measured on ordinal scales. Therefore, traditional reliability analysis procedures want to be properly adjusted.

Directors
  • Giuseppe Boari, Università Cattolica del Sacro Cuore
  • Marika Vezzoli, University of Brescia
Checked Open Submissions Checked Peer Reviewed

SOLI-M3.3 - Compositional analysis

Compositional data are data where the elements of the composition are non-negative and sum to unity. Geometrically, compositional data with J components has a sample space of the regular unit J-simplex. In this session, the problems of robustness and the presence of the zeros in the dataset are studied when latent variables are extracted from this kind of data. Of course, the properties of compositions must be strictly taken into account before any attempt of interpretation.

Directors
  • Michele Gallo, Department of Human and Social Sciences, University of Naples L'Orientale
  • Rosaria Lombardo, Second University of Naples
  • Marica Manisera, University of Brescia
Checked Open Submissions Checked Peer Reviewed

SOLI-A1.1 - Bayesian Models in Economics and Finance

In this session we focus on new Bayesian modeling time series with applications in economics and finance.

Directors
  • Francesco Lagona, University Roma Tre
  • Marica Manisera, University of Brescia
  • Antonello Maruotti, Libera Università Maria Ss. Assunta University of Southampton
  • lea petrella
Checked Open Submissions Checked Peer Reviewed

SOLI-A1.2 - Space and space-time models B: methods and environmental applications

Hierarchical and structural equation models for spatiotemporal data are based on spatial and temporal trends which are defined by latent components. Session topics are
1) the INLA approach for Gaussian spatiotemporal large data sets;
2) the mixed model representation of functional data with application to spatiotemporal series of atmospheric vertical profiles.
3) Hierarchical spatiotemporal for air quality high resolution forecasting.
4) spatiotemporal SEM are considered for the non Gaussian case.

Directors
  • Alessandro Fasso, University of Bergamo
  • Silvia Golia, University of Brescia
  • Giovanna Jona Lasinio, Department of Statistical Sciences "Sapienza" University of Rome
Checked Open Submissions Checked Peer Reviewed

SOLI-A1.3 - Advances in Rasch Analysis

The papers that will be presented in this session refer to the family of Rasch models. These models are able to measure latent variables from data coming from surveys and have found applications in many fields, including education, psychology, health and marketing. This session aims to discuss recent developments in Rasch analysis.

Directors
  • Maurizio Carpita, University of Brescia
  • Silvia Golia, University of Brescia
Checked Open Submissions Checked Peer Reviewed

SOLI-A1.4 - Unobservable features detection in finance

This session covers some main topics in the context of the statistical techniques aimed at identifying latent structures, modelling latent variables and measuring latent features within the financial research field. The interest will be focused on volatility forecast, risk measurement, evaluation of the dynamic association between assets.

Directors
  • Maurizio Carpita, University of Brescia
  • Giovanni De Luca, University of Naples Parthenope
  • Paola Zuccolotto, University of Brescia
Checked Open Submissions Checked Peer Reviewed

SOLI-A1.5 - Advances in longitudinal data analysis

The session will include new modelling approaches to the analysis of longitudinal data with complex latent dependence structures.

Directors
  • Francesco Lagona, University Roma Tre
  • Marica Manisera, University of Brescia
  • Francesca Martella, Dipartimento di Scienze Statistiche
Checked Open Submissions Checked Peer Reviewed

SOLI-A1.6 - Space and space-time models C: methods and applications

Hierarchical and structural equation models for spatiotemporal data are based on spatial and temporal trends which are defined by latent components. Session topics are
1) space-time models for European air pollution exposure distribution and risk indexes evaluation;
2) multivariate geostatistical skew-normal models;
3) Simultaneous diagonalization of the sample matrix variogram for multivariate spatiotemporal processes.

Directors
  • Alessandro Fasso, University of Bergamo
  • Silvia Golia, University of Brescia
  • Luigi Ippoliti, University of Chieti-Pescara
Checked Open Submissions Checked Peer Reviewed

SOLI-A2.1 - Latent variables in time series

The scope of this session is to assess to what extent and how  latent variable models might support the analysis of time series analysis. The session will provide  contributions on the most recent trends, innovative approaches and future challenges in the field. In particular, the interest will be focused on the identification of latent structures and the modeling of latent variables in economic and financial time series.

Directors
  • Cira Perna, University of Salerno
  • Paola Zuccolotto, University of Brescia
Checked Open Submissions Checked Peer Reviewed

SOLI-A2.2 - Job satisfaction and quality of work: measures and models

This session aims at discussing measures and models for some relevant variables (job satisfaction, quality of work and well-being) which are unobservable but have an important effect for assessing individual and collective behavior. Some papers mainly focusses on the problem of the correct definition and measure of such latent variables whereas others show how innovative models are able to capture the main features of such phenomena. All the works debate on real data coming from official surveys related to the Italian system.

Directors
  • Domenico Piccolo, University of Naples Federico II
  • Paola Zuccolotto, University of Brescia
Checked Open Submissions Checked Peer Reviewed

SOLI-A3.1 - Latent traits for Educational Models

Many studies on the educational systems are focused on the difference in scholastic achievement. Several models are available for the analysis of these data. Most of them fall into the category of added value models. However these models do not often take in account many important aspects as the longitudinal and multilevel structure of data, the presence of different clusters of students, the equal discriminating power of the test items. The family of Latent traits models can overcome these methodological problems.

Directors
  • Giorgio Vittadini, Università di Milano Bicocca
  • Paola Zuccolotto, University of Brescia
Checked Open Submissions Checked Peer Reviewed


Contact: sis2013@eco.unibs.it