Working meeting "Handling categorical and continuous data" - May, 16th 2014

Organizers: Nathalie Villa-Vialaneix (INRA) and Sébastien Da Veiga (Snecma)

It will take place on:

May 16th 2014, Amphithéâtre Hermite, Institut Henri Poincaré, Paris.

Presentation of the workshop

The goal of this workshop is to give an overview of statistics and optimization methods to handle problems with both categorical and continuous input or output variables. Many practical applications now involve such variables and practitioners have to focus on dedicated tools.
For completeness, the day will be divided into three sessions: Clustering, Optimization and Regression. In each session, standard tools and recent advances will be presented. Further note that we will try to give “live” demonstrations of software packages or lists of available softwares, whenever possible.

Preliminary Program

9:15 : Welcome of the participants

9:30 – 9:45 : Introduction – S. Da Veiga (Snecma) - slides

9:45 – 11:00 : Clustering I – Marie Chavent (Université Lille 1 & INRIA) (Université Bordeaux 2 & INRIA) – Variable Clustering and Mixed Data. The ClustOfVar package - slides

11:00 – 11:15 : Coffee Break

11:15 – 12:30 : Clustering II – Matthieu Marbac (Université Lille 1 & INRIA) – Model-based clustering of Gaussian copulas for mixed data - slides script R

12:30-14:00: Lunch break

14:00 – 14:45 Optimization I – Claudia D'Ambrosio & Luca Mencarelli (LIX Ecole Polytechnique) – Introduction to Mixed Integer Nonlinear Programming - slides

14:45 – 15:30: Optimization II – Claire Lizon (IFPEN & LIX Ecole Polytechnique) – Mixed Integer Nonlinear Optimization for Well Placement - slides

15:30 – 15:45: Coffee Break

15:45 - 17:00: Regression – Claire Cannamela (CEA) – Metamodels involving both categorical and continuous inputs variables - application to a spectrometry gamma problem - slides

17:00 : End of the workshop

Registration is free but appreciated.


CNRS