Mehdi Dagdoug

Assistant Professor of Statistics

Department of Mathematics and Statistics
McGill University

• Survey sampling
• Missing data
• Statistical learning
• High-dimensional inference

I am interested in statistical inference from survey data, both for finite population quantities and for superpopulation quantities, particularly in difficult settings: a large number of auxiliary variables, domains containing few sampled units, or nonresponse. Survey data are often neither independent nor identically distributed. I aim to understand how classical methods behave in this setting and how new tools, including machine learning, can improve them, by studying their asymptotic properties, how their uncertainty can be quantified, and when they are optimal.

This leads me to work on machine learning for finite population inference, statistical learning from survey data, missing data and high-dimensional inference.

mehdi.dagdoug@mcgill.ca · Burnside Hall, room 1230

Contact Research Publications CV (PDF)