Modelling non-linear data with Generalized Additive Models (GAMs), using the mgcv package with Chris Mainey
We are very excited to announce that the NHS-R Community will be hosting its annual conference, including pre-conference online talks, online workshops and the main conference, during October 2023.
The online conference workshop will be hosted via Zoom and take place on the following date:
This workshop is a follow-on from my talk at last year’s NHS-R conference on modelling ‘wiggly’ data using Generalized Additive Models. Regression models are a great approach for modelling relationships, but they make certain assumptions about linear relationships that don’t always hold true in real-world datasets.
The workshop is aimed at people who have some experience of regression modelling and want to take it further, learning a more advanced and flexible class of modelling that also applies to non-linear data.
This workshop will introduce the class of models (explaining a GAM as a GLM with smooth functions of predictors). We will then walk through a simple healthcare based example of a GAM. We will discuss how to interpret the GAM, assess whether the fit is good, and look at ways to visualise the smooth functions. We will then take a detour through what the smoothers actually do, using piece-wise cubic splines, as our example.
We will finish the workshop with a more complicated example, asking attendees to fit the model, examine the smoothers and answer a couple of questions about them.
Please be advised that this course is an advanced regression modelling course, and it is essential that attendees have some experience of fitting regression models in R to understand the application of GAMs.
By registering for this event, you will qualify for the following:
A Zoom link will be provided to you to attend the conference workshop virtually. Please note that if subtitles are used, they are likely to be generated by Zoom thus, may encounter some issues converting technical language and acronyms to text.
If you require any additional information or have any accessibility requests, please do not hesitate to contact the Conference Team via email@example.com.
Additional Conference Related Activities
Please note that the conference will also include the following separate activities:
NHS Data Scientist
Chris Mainey is an NHS data scientist and leads the specialist analytics team at North Central London ICB. He has experience in patient safety at NHSE, national benchmarking at University Hospitals Birmingham, and in public health, clinical networks, acute trust informatics, and leading analytical and data science teams. He completed his PhD in applied statistics and machine learning at UCL in 2020, focussing on methods for analysing incident reporting data in the NHS using multi-level regression models, Generalized Additive Models (GAMs), tree and random forest methods and natural language processing (NLP). He has a strong interest in training and teaching, and is a senior fellow of the NHS-R community.