Generalised Additive Modelling in statistics

Organisator(en) Márton Sóskuthy
Institution(en) University of British Columbia
Veranstaltungsbeginn 28.04.2022, 15.30 Uhr
Veranstaltungsende 28.04.2022, 17.30 Uhr

Description

This workshop focuses on Generalised Additive Modelling, a regression modelling technique for the flexible analysis of non-linear patterns. GAMs are becoming standard in the analysis of a wide range of different types of linguistic data that show temporal or spatial dependencies: formant contours, eye-tracking data, EEG measurements, midsagittal tongue surface data, map data, etc.  We will take an approach that balances GAM theory with practical coding exercises. The goal of the workshop is to allow you to set up, run and interpret your own GAM models as well as critically evaluate those of others.


The workshop will be run over four two-hour sessions, each of which will comprise a theoretical discussion and a practical exercise. The planned schedule is as follows:

April 28 (Thurs), 3.30-5.30pm
What is a generalised additive model? When should I use one? Fitting simple GAMs.

May 19 (Thurs), 3.30-5.30pm
GAMs for data with multiple trajectories: autocorrelation, AR error models and generalised additive mixed models.

June 8 (Wed), 3.30-5.30pm
GAMs with non-linear interactions (e.g. how tongue contours change over time; how formant contours shift with sound change). Advanced GAM visualisation with custom plots & videos.

June 22 (Wed), 3.30-5.30pm
The goal of this session is to provide personalised advice on Generalised Additive Modelling based on your specific needs. This could be a session on a more specialised topic that we don't cover in earlier sessions, or a data-surgery type session where we analyse one or two data sets provided by the group.

Please fill in this registration form if you are planning to attend: https://forms.office.com/r/L50SdjgBGT

For access information please contact zygis@leibniz-zas.de.