EpiNow2

EpiNow2 copied to clipboard

Add support for time-varying ascertainment

5

As the title. On a single data source this could/would only be in the form of a user input but if any additional complexity (i.e. multiple data sources) it could...

seabbs

Improve discretise

2

`primarycensoreddist` now supports numerically solved double censored pmf calculation (https://primarycensoreddist.epinowcast.org/dev/articles/primarycensoreddist.html#compute-the-primary-event-censored-probability-mass-function-pmf-with-dprimarycensoreddist) which we should replace our current approximate `dicretise` approach with either wrapped or directly.

seabbs

Improve internal handling of PMFs in stan

1

`primarycensoreddist` now supports efficient numerically solved correctly double censored stan PMFs (https://github.com/epinowcast/primarycensoreddist/blob/de62b580e9089a31311981ad2c1c76f477238865/inst/stan/primary_censored_dist.stan#L420). It will shortly support analytical solutions for common uses cases. See here for proposed approaches to working with...

seabbs

The growth rate is currently estimated via the log derivative of infections, see https://github.com/epiforecasts/EpiNow2/blob/35483d03327cdcce4be0a3ecf5aa7c9466aa42be/inst/stan/functions/generated_quantities.stan#L33-L39 I know that this is the classical interpretation of the growth rate, but wanted to point...

adrian-lison

Locally stationary GPs

1

Extend the current GP functionality to be locally stationary by using a patch based approach of stationary Gps with the GP parameters being related with hierarchical distributions. The main issue...

seabbs

ARMA spectral density

1

As suggested by @SamuelBrand1 ARMA models have a spectral density (https://en.wikipedia.org/wiki/Autoregressive_moving-average_model) this means we should be able to implement approximate AMRA models within our GP framework. This builds off discussions...

seabbs

The stan R package community is adopting the use of algorithms like `pathfinder` to initialise their models. See the workflow in https://github.com/epinowcast/epinowcast/pull/504. Would be good to join the party.

jamesmbaazam

Document all stan code

1

The stan code here has very little documentation, so there is a need to make this a priority to improve navigation. This can be done with Doxygen as mentioned here...

jamesmbaazam

Change Getting started documentation to point users to faster models

5

This issue can be solved after merging the benchmarking vignette in #695. I was curious about how the paper ["Real-time estimation of the epidemic reproduction number: Scoping review of the...

jamesmbaazam

Vignette on `estimate_infections()` examples has errors in code chunks

2

The example on passing a truncation distribution to `estimate_infections()` errors in the vignette [here](https://epiforecasts.io/EpiNow2/articles/estimate_infections_options.html#adjusting-for-truncation-of-the-most-recent-data). As an aside, should we generally ensure that the vignettes' model specifications do not produce divergences?

jamesmbaazam