Statistics Seminar - Predicting under intervention: the intersection of causal inference and prediction
| Dates: | 23 October 2024 |
|---|
| Times: | 14:00 - 14:00 |
|---|
| What is it: | Seminar |
|---|
| Organiser: | Department of Mathematics |
|---|
| Who is it for: | University staff, External researchers, Current University students |
|---|
| Speaker: | Matthew Sperrin |
|---|
|
Abstract:
Prediction and causal inference are usually seen as separate tasks (although sometimes people are somewhat confused about which they are doing). In prediction, we typically wish to forecast an outcome as accurately as possible and may be less interested about understanding such a prediction. In causal inference, we wish to estimate the effect of an intervention.
Often, however, causal inference is useful in prediction. It's useful directly when we wish to 'predict under intervention' (what is the risk of Y if I do A). It's useful indirectly when we want to build models that can be generalised or transported more reliably.
In this talk I'll give some methods and applications related to current work on these topics - including the development of a model to predict future cardiovascular risk 'under intervention'.
Speaker
Matthew Sperrin
Organisation: Manchester
Travel and Contact Information
Find event
Frank Adams 2
Alan Turing Building
Manchester
