We return to another case study from the short course Simulation-Based Inference for Epidemiological Dynamics (SBIED). Chapter 17 is Lesson 5 of SBIED. In Chapters 11–14, we used a single measles outbreak as a relatively simple example to demonstrate POMP models and inference. Now we see how an extension of this analysis becomes a topic of scientific interest. Modeling longer time series can be more challenging, since it is not simply enough to have a model that can describe how an epidemic wave surges and then retreats. On the other hand, the additional data from a sequence of outbreaks can inform a more detailed model.
We will focus on two new topics arising in the case study:
When can we make a causal interpretation of estimated model parameters?
We see that the scale of the variability matters for successful modeling. We show how to model dynamic stochasticity by adding noise to the rates of a Markov chain.
| Slides | ||
| Annotated slides | ||
| Notes | ||
| R script | R | |
| Recording, Chapter 17, Part I | Model development | (40 min) |
| Recording, Chapter 17, Part II | Interpreting the results | (40 min) |
| Model construction and analysis script | Rnw | |
| Supplement: profile likelihood calculation | HTML |
Back to course homepage
Acknowledgements
Source code
for these notes