Instantiate CFA's Epidemia-For-Flu Model Via MSR

[AFg6K7h4fhy2]

Goal

Verify that DHM's influenza forecasting repository cfa-forecast-renewal-epidemia, which utilizes epidemia, can be instantiated via MSR.

Context

Utility from the multisignal-epi-inference repository is derived from being able to modularly construct renewal processes. One renewal model that should be able to be specified via MSR tools is the aforementioned cfa-forecast-renewal-epidemia. This port-over process, in expectation, might allow MSR developments to identify structural inadequacies or missing features and or usage capabilities.

Required Features

• Create quarto walk-through detailing the port-over steps and providing documentation.
• Validate MSR implemented cfa-forecast-renewal-epidemia via simulation and fitting on NHSN data, which is the data cfa-forecast-renewal-epidemia has historically been used on.

Out Of Scope

• Non-epidemia related MSR-implementations of renewal processes.
• epidemia beyond the scope of what's captured in cfa-forecast-renewal-epidemia.

Related documents

• Epidemia, website
• Epidemia, GitHub
• Epidemia, installation
• Epidemia, model description
• Epidemia, model implementation

[seabbs]

DHM's implementation of Epidemia.

I interpreted this as @dylanhmorris implementation of the tool epidemia but I know the meaning is the use of epidemia for flu forecasting. I imagine everyone is on the same page but perhaps worth a reword for clarity.

Making this point as reimplementing all of epidemia is a much bigger task than the quite simple model that ended up being used for flu.

Validate MSR implemented Epidemia model with NHSN data

It seems like if seeking to duplicate a new implementation should be validated against the old one rather than just against old data (i.e via simulation)

[AFg6K7h4fhy2]

@seabbs

Thank you for the feedback, Sam. I've edited the issue. Wrt your first point, I think [that the target is a private repository] affected my wording, i.e I tried to cloak cfa-forecast-renewal-epidemia. Wrt the second point, this was simply a lack of detail on my end.

[AFg6K7h4fhy2]

After some conversation {DHM, DamonB, GeorgeVY, me}, it has been decided, with moderate strength, to contain this port-over-implementation in the tutorials folder "for the time being".

[AFg6K7h4fhy2]

The model has, more or less, been instantiated; it can be found here https://github.com/CDCgov/pyrenew-flu-light. Evaluate and further refinement in progress, see this issue.