2-Non-stationary processes, in discrete and continuous time

2-4 May, 2012

19 avenue du Maine
75015 Paris

Speakers :

  • Konstantinos Fokianos, Chypre
  • José Rafael Leon, UC Venezuela
  • Marc Lavielle, Inria Saclay
  • Dag Tjøstheim, Univ. of Bergen




Wed May 2

Thu May 3

Fri May 4

Wed May 9


ENGREF, 19 avenue du Maine  75015 Paris

UCP, Site St Martin, AGM

E-554 (5th floor)

UCP, Site St Martin, AGM

E-554 (5th floor)

UCP, Site St Martin, AGM

E-554 (5th floor)

10-12/13 am

M. Lavielle


K. Fokianos


D. Tjostheim

K. Fokianos

J.R. Leon

2-4/5:30 pm

M. Lavielle


J.R. Leon

J.R. Leon


Titles and abstracts:

  • Konstantinos Fokianos : "Statistical Analysis of Count Times Series"

Télécharger le fichier «Fokianos_may2012.pdf» (74.9 KB)

Documents de cours :

Télécharger le fichier «Fokianos_CTSM.pdf» (196.4 KB)

- Télécharger le fichier «Fokianos_JMVA.pdf» (752.6 KB)

- Télécharger le fichier «Fokianos_NLPA.pdf» (636.3 KB)

- Télécharger le fichier «Fokianos_PA.pdf» (278.3 KB)

- Télécharger le fichier «Fokianos_STAPRO.pdf» (384.3 KB)

  • José Rafael Leon : "Chaos de Ito-Wiener and its applications in statistics of processes"

In this course we will consider parametric and nonparametric estimation for processes that can be built by using some nonlinear functionals of Gaussian processes. We shall consider no only processes of real parameter but also random fields. The examples that we have in mind are: fractional diffusions, non isotropic Gaussian random fields and the estimation of their anisotropy, and estimation via functionals of level sets

Documents de cours :

- Télécharger le fichier «Leon1_may2012.pdf» (1.1 MB)

- Télécharger le fichier «Leon2_may2012.pdf» (490 KB)

  • Marc Lavielle : “Drug and desease modelling: models and methods”

Model based-drug development (MBDD) is accepted as a vital approach in understanding patient risk/benefit and attrition. At the core of MBDD lies Modelling and Simulation (M&S), a technology providing the basis for informed, quantitative decision-making.

M&S facilitates the continuous integration of available information related to a drug or disease into constantly-evolving mathematical models capable of describing and predicting the behaviour of studied systems to address the questions researchers, regulators, and public health care bodies face when bringing drugs to patients.

The mathematical models used for describing complex biological phenomena are complex: they are usually be based on (Ordinary, Partial, Stochastic...) Differential Equations. Furthermore, the statistical component of the model should be carefully taken into account, in order to properly describe the variability between patients of the response to a same treatment. Because of this complexity, new methods need to developed for parameter estimation, model assessment, model selection,...

I will present several  models  (Pharmacokinetic/Pharmacodynamic, HIV dynamics, epilepsy ...) and some statistical tools that we have developed for these models and which are now widely used for practical applications.

  • Dag Tjøstheim : "A local correlation coefficient"

Nonlinear estimation in a nonstationary environment

The talk consists of two parts. In the first part I will give an overview
over nonparametric estimation and testing in nonstationary Markov time series using the splitting technique. The relationship to nonlinear cointegrating regression will be pointed out. Unlike the stationary case there could be a fundamental difference between regression and autoregression. In the second part I will look at some new results for parametric estimation for
nonstationary time series. Again the Markov splitting technique is useful. This approach will be compared to an approach using local time.

Télécharger le fichier «Tjostheim_may2012.pdf» (368.1 KB)

Inscription :

Contact : thomas.ballesteros@u-cergy.fr ( thomas.ballesteros @ u-cergy.fr)

Une table ronde sera organisée le 11 mai 2012 sur Paris pour exploiter les notions induites lors des cours dans une dynamique de recherche tant théorique qu’appliquée.