Copula-based regression models: the good, the bad and the ugly
Bouchra Nasri, PhD
Assistant Professor | Department of Social and Preventive Medicine |
Université de Montréal
WHERE: In-Person - 2001 ÎÛÎÛ²ÝÝ®ÊÓƵ College Avenue, Rm 1140;
NOTE: Social gathering (Kitchen Lounge Rm 1205)Ìý 3-3:30pm
Abstract
Copula-based regression models have become very popular in several fields of applications over the past decade, including in health sciences. While these approaches are interesting, especially for capturing complex dependence structures, they are often used in a restricted or mistaken way in practice. In this talk, I will present several copula-based regression models and discuss their advantages and disadvantages.
Speaker Bio
Professor Nasri is a faculty member of Biostatistics in the Department of Social and Preventive Medicine. Prof. Nasri is an FRQS Junior 1 Scholar in Artificial Intelligence in Health and Digital Health. She holds an NSERC Discovery Grant in Statistics for time series dependence modelling for complex data. Prof. Nasri and her team are working mainly on the development of statistical learning methods, artificial intelligence methods and mathematical models for infectious diseases and public health threats related to climate change. Prof Nasri completed a Ph.D. under the supervision of Taoufik Bouezmarni, Taha Ouarda, and André St-Hilaire, an MSc under the supervision of Salah El-Adlouni and Taha Ouarda at INRS-ETE on developing Statistical approaches for extremes related to Climate Changes and an Ingénieur d'État degree (eq. MSc) in Statistics at INSEA. She also completed a postdoctoral fellowship in theoretical statistics at ÎÛÎÛ²ÝÝ®ÊÓƵ and HEC Montréal funded by the FRQNT, CANSSI and GERAD. She authored and co-authored several papers on time series, spatial dependence, multivariate statistics, compartmental modelling, text mining, and evidence synthesis.
Website:
Ìý