Event

Nema Dean (University of Glasgow)

Friday, March 29, 2019 13:00to16:30
McIntyre Medical Building Room 521, 3655 promenade Sir William Osler, Montreal, QC, H3G 1Y6, CA

Title: Introduction to Statistical Network Analysis

Abstract:

Classical statistics often makes assumptions about conditional independence in order to fit models but in the modern world connectivity is key. Nowadays we need to account for many dependencies and sometimes the associations and dependencies themselves are the key items of interest e.g. how do we predict conflict between countries, how can we use friendships between school children to choose the best groups for study tips/help, how does the pattern of needle-sharing among partners correlate to HIV transmission and where interventions can best be made. Basically any type of study where we are interested in connections or associations between pairs of actors, be they people, companies, countries or anything else, we are looking at a network analysis. The methods falling under this area are collectively known as “Statistical Network Analysis” or sometimes “Social Network Analysis” (which can be a bit misleading as we are not only talking about Facebook and the like). This workshop will give a general introduction to networks, their visualisation, summary measures and statistical models that can be used to analyse them. The practical component will be in R and attendees will get the most benefit if they are able to bring a laptop along to work through examples.

Speaker

Nema Dean is a Senior Lecturer of Statistics in the School of Mathematicss and Statistics at the University of Glasgow. Her research interests are in developing new clustering and classification methods. Past work has involved research on finite mixture model based methods and variations that incorporate variable selection and semi-supervised updating. Currently she is working on creating hybrid clustering methods using both parametric and classical algorithmic approaches. She have also developed new mixture model clustering methods for discrete and space-restricted data.

Follow us on

Back to top