Model-based clustering and classification permit to discover structure in complex data and to assign observations to latent groups. These methods are widely used in statistics and machine learning, with applications ranging from social sciences to genomics.

This master class will introduce the main ideas underlying model-based clustering and classification, with an emphasis on mixture models, likelihood-based inference with the EM algorithm, and probabilistic interpretation of clusters and classes.

The course will be primarily lecture-based, interleaved with illustrative code demonstrations in R.

The organization of this master class was possible thanks to the PHC Ulysses grant INTEGRATE. 

Brendan is Professor of Statistics in University College Dublin, Ireland and he works on statistical modelling in a wide range of domains, including agri-food, biomedical and social sciences. He has recently served as head of the School of Mathematics & Statistics at University College Dublin. He has been a visiting fellow at the Center for Statistics & Social Sciences at University of Washington and Institut d’Études Avancées de Lyon. He has served as Area Editor of the Annals of Applied Statistics and he is currently a member of the Statistical and Methodology Review Committee for Proceedings of the National Academy of Sciences. He was recently elected as Fellow of the American Statistical Association and has been selected as the Honorary Officer for Journals by the Royal Statistical Society.

The course will be held on March 19, 2026, from 9:30 a.m. to 1:00 p.m, in room C1.1.01 at AgroParisTech, 22 Place de l'agronomie, 91120 Palaiseau.

Pre-requisites: Statistics, linear models. Familiarity with R would help.

Registration is free but mandatory.

Due to the limited number of seats, registration is subject to validation by the organizers.