This talk will feature Chris Mainey discussing the use of ‘Generalized Additive Models (GAMs)’ to fit regression models to ‘wiggly’ or non-linear data. Sometimes the relationships we are interested in are not a simple straight-line. Polynomial terms and combining them to for ‘splines’ will be introduced and then used in GAMs. This will be demonstrated in R as part of the talk. This will be an exciting talk for those wanting to deepen their knowledge of regression.
For in-person attendees, the talk will be followed by pizza/salad and soft drinks, thanks to our sponsors: the R Consortium.
Resources
- GAM workshop for NHS-R Community Conference 2023 GitHub repo: https://github.com/chrismainey/GAMworkshop
- Slides: https://chrismainey.github.io/GAMworkshop/GAMworkshop.html
About the speaker
Chris Mainey is Head Data Scientist for NHS Birmingham and Solihull ICB. He has led analytics and data science teams in areas including hospital benchmarking, population health management, NHS commissioning, mortality statistics and patient safety. He completed his PhD in Applied Statistics and Machine Learning at UCL, focussing on modelling count data and applying natural language processing to NHS datasets. He is a Senior Fellow of the NHS-R Community and is keen on teaching and spreading data science skills, coding, statistics, and good analytical practise. He’s very fond of a good coffee, GAM models, and putting code on GitHub.
Those interested in the NHS-R community may also want to see:
- NHS-R Community – Promoting the use of R in the NHS (nhsrcommunity.com)- open to researchers and anyone in public sector/health related fields.Join the Slack: https://nhsrcommunity.slack.com/
- GitHub: https://github.com/nhs-r-community
- NHS-R Conference material (more content published soon): https://www.youtube.com/channel/UCMwM-3tg_-Pbx8hKO78q5EA
Photos
Citation
@online{mainey2023,
author = {Mainey, Chris},
title = {Fitting {Wiggly} {Data:} {Generalized} {Additive} {Models}},
date = {2023-11-26},
url = {https://warwickrug.github.io/posts/2023-11-23-fitting-wiggly-data},
langid = {en}
}