5/07/2021 (Online)
Short course on "Modeling Survival Outcomes with High Dimensional Predictors: Methods and Applications"
by Yi Li (University of Michigan)

The EDT StatActu, together with KULeuven and the ARC “Imperfect data: From Mathematical Foundations to Applications in Life Sciences” is organising thsis event.

In the era of precision medicine, survival outcome data with high-throughput predictors are routinely collected. These high dimensional data defy classical survival regression models, which are either infeasible to fit or likely to incur low predictability because of overfitting. This short course will introduce various cutting-edge methods that handle survival outcome data with high dimensional predictors. I will cover statistical principles and concepts behind the methods, and will also discuss their applications to the real medical examples.

Time permitting, I intent to cover the following topics.
1.      Survival analysis overview: basic concepts and models, e.g. Cox, Accelerated Failure Time (AFT), and Censored Quantile Regression (CQR) Models;
2.      Survival models with high dimensional predictors (p>n): Regularized methods and Dantzig selector;
3.      Survival analysis with ultra-high dimensional predictors (p>>n): Screening Methods, e.g, Principled sure independent screening (PSIS), Conditional screening, IPOD, Forward selection, etc;
4.      Inference for survival models with high dimensional predictors (p>n).
 Audience only needs to have some basic knowledge of regression analysis and survival analysis. The relevant papers and software for this short course can be found in: http://www-personal.umich.edu/~yili/resindex.html.
 

10-11/12/2019
Short course in Actuarial science on "Longevity, pension and long term care"
by Massimiliano Menzetti, Universita della Calabria

In recent decades, most developed countries have been characterized by significant improvements in longevity. In 2012, the International Monetary Fund affirmed that longevity risk was a major risk in developed countries due to its impact on social security and healthcare systems. Otherwise longevity risk also affects private organisations that offer whole life annuity benefit as insurance company and occupational pension schemes.
Many stochastic mortality models have been developed to measure longevity risk in national population but a potential divergence could be observed between the national population’s mortality and the pensioners’ mortality (basis risk). Moreover, when dealing with private pension schemes, the records available are often referred to a few years and there may be missing data.

The first goal of this course is to focus on model implication in representing the longevity risk in social security systems and in private pension schemes.
The focus will be on two aspects: 
1) modelling mortality divergences produced by different socio-economic characteristics of the insured population respect to the national one (basis risk),
2) Modelling the mortality evolution of a small population with respect to the national population’s mortality.

As previously stated, increasing longevity has also implications on health expenditure.
An older population is more exposed to morbidity and disability.
Specifically, very few models have been developed to represent the joint evolution of longevity and disability risk.
The second goal of the course will be on a proposal of a joint projection model for longevity and disability and its implication in hedging strategies.

 

11-12-13/02/2020
Short course on "Modélisation individuelle des réserves en assurance non-vie"
by Mathieu Pigeon (Université du Québec à Montréal)

Le short course a pour objectif principal d’initier les étudiantes et les étudiants aux nouvelles approches en modélisation des réserves en assurance non-vie. Un rappel des principales approches collectives sera effectué mais le cours se focalisera sur les approches individuelles (ou granulaires, ou micro, en fonction de la terminologie utilisée).
La dynamique individuelle du développement des sinistres sera présentée et on abordera plus particulièrement (1) les modèles basés sur des processus (de Poisson) avec marqueurs, (2) les modèles pour les situations anormales (robustesse) et (3) l’utilisation des techniques liées à l’apprentissage statistique. Une place importante sera accordée à différents aspects « pratiques » tels que l’analyse de bases de données, la sélection des variables explicatives pertinentes, la mise en oeuvre informatique, etc. Les applications pratiques se feront avec le logiciel R bien qu’une connaissance approfondie de ce logiciel ne soit pas nécessaire pour suivre le cours.

09, 11/03/2020
Short course on "Data Stream and Monitoring"
by Ansgar Steland, RWTH Aachen, Germany

Data sets and data streams, i.e. time series of variables of interest, are ubiquitous but often affected by changes (breaks) where important distributional features - most prominently the mean - change. Such change-points may be highly informative. For example, they may indicate a fault in a production line, unexpected news or events increasing financial risk or a critical event calling for action in  intensive care monitoring. Compared to hypothesis testing, the main difference is the fact that one wants to apply a statistical decision procedure either on-line at each time instant when a new observation arrives to trigger signals, or off-line (retrospective) to all successive subsamples to figure out whether a change is present, whereas a hypothesis test is applied once to the full data set. Therefore, such procedures can be designed to control, as usual, error probabilities, but it is also common to control the average number of observations (called average run length, ARL) until a false or right signal is given. From a theoretical viewpoint, recalling that estimators and test statistics are typically asymptotically given by averages (of transformed observations), statistics has now to deal with the whole sequence of such averages and not only with one average. This leads to partial sum processes and their properties. This approach is crucial to construct and study procedures for dependent data, both when relying on parametric models or working in a nonparametric framework. 

This short course gives an introduction and overview of some commonly used procedures, their statistical design, and some approaches to derive asymptotic properties.

1. Introduction
2. Basic Procedures
3. ARL approximations
4. Asymptotics

09-10/11/2020 (Online)
Short course on "Causality: Models, Learning, and Invariance"
by Jonas Peters, University of Kopenhagen

In science, we often want to understand how a system reacts under interventions (e.g., under gene knock-out experiments or a change of  policy). These questions go beyond statistical dependences and can therefore not be answered by standard regression or classification techniques. In this tutorial we will learn about the powerful language of causality and recent developments in the field. No prior knowledge about causality is required. More precisely, we introduce structural causal models and formalize interventional distributions. We define causal effects and show how to compute them if the causal structure is known. We discuss assumptions under which causal structure becomes identifiable from observational (and interventional) data and describe corresponding methodology. If time allows, we present connections between causality and distributional robustness.

 

04, 11, 18/02/2019
Short course on : "Introduction to high-dimensional statistics"
by Eugen Pircalabelu, ISBA - UCLouvain

08-09/04/2019
Short course on : "From frontier models in production econometrics to extreme value theory in risk handling"
by Abdelaati Daouia, Toulouse School of Economics, at ULB

 

 

01, 08, 11/12/22017
Short course on "An introduction into the asymptotic theory of statistical experiments"
by Marc Hallin (ULB)

15-16/02/2018
Short course on : "Computational aspects of regularisation methods for high-dimensional data problems"
by Sylvain Sardy (Université de Geneve)

05, 07, 12, 14/03/2018
Short course on : "Introduction into high-dimensional statistics : Theory and Practice"
by Johannes Lederer (Bochum University)

17-18/05/2018
Short Course on : "Survival analysis methods for micro-level claim reserving"
by Olivier Lopez (Université Pierre et Marie Curie Paris VI)

 

14-15/11/2016
Short course on "Big data in insurance"
by Olivier Lopez (Université Pierre et Marie Curie Paris VI)

 

19-20/11/2015
Short course on "Nonlinear Valuation and XVA under Credit Risk, Collateral Margins and Funding Costs"
by Damiano Brigo (Imperial College London)

14-15/03/2016
Short course on "PLS regression methods and extended tools with application to –omics data"
by Philippe Bastien (L’oréal Research and Innovation, France)

 

 

17/09/2014
Short course on "Analysis of competing risks data for inference and prediction"
by Alina Nicolaie (Universiteit Gent)

14-15/10/2014
Short course on "Cure models and their estimation methods and applications"
by Yingwei Peng (Queen's University, Kingston)

14-15/10/2014
Short course on "The spanish pension system reforms in 2011 and 2013"
by Beatriz Rosado Cebrián (University of Extremadura)

30-31/03/2015
Short course on "Change Point Analysis"
by Claudia Kirch (Karlsruhe Institute of Technology)

26-27/05/2015
Short course on "OPTIMAL STOPPING – Theory and Applications "
by F. Thomas Bruss (ULB)

 

05/11/2013
Short course on "Dimension reduction in regression"
by François Portier (UCL)

21/11/2013
Short course on "Using Machine Learning and Bayesian Methods to Analyze Large Supernovae Datasets"
by Melvin Varughese (University of Cape Town)

06/05/2014
Short course on “An introduction to multivariate and dynamic risk measures”
by Arthur Charpentier (UQAM)

19-20/05/2014
Short course on “An Introduction to the Joint Modeling of Longitudinal and Survival Data, with Applications in R”
by Dimitris Rizopoulos (Erasmus University Medical Center, the Netherlands)

 

26, 28/09/2012 and 02/10/2012
Short course on "An Introduction to the Bootstrap: Panacea for Statistical Inference"
by Cédric Heuchenne (Université de Liège)

21-22/05/2013
Short course on "Copula-based Dependence Models"
by Elif Acar (University of Manitoba)

 

15/12/2011
Short course on “An introduction to weak dependence techniques”
by Paul Doukhan (University of Cergy-Pontoise)

02-03/05/2012
Short course on “Dimension reduction using single-index regression models”
by Olivier Lopez (Paris VI University)

11/05/2012
Short course on “Numerical methods in income protection models”
by Isabel Ferraz Cordeiro (Universidade do Minho, Braga and CEMAPRE, Lisbon)

29-30/05/2012
Short course on “Empirical process theory for statistics”
by Jon A. Wellner (University of Washington)