Add SIR model with unknown time of initial infection #2460
Merged
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
38 tasks
|
@fehiepsi would you mind reviewing the |
pyro/contrib/epidemiology/sir.py
Outdated
| self.pre_window = pre_window | ||
| self.post_window = len(data) | ||
|
|
||
| # Expect a single external infection during the pre_window. |
There was a problem hiding this comment.
can you please explain this assumption?
| # after observations were made (in post_window). | ||
| rho0 = pyro.sample("rho0", dist.Uniform(0, 1)) | ||
| rho1 = pyro.sample("rho1", dist.Uniform(0, 1)) | ||
| rho = torch.cat([ |
There was a problem hiding this comment.
comment on shape of rho?
| S2I = pyro.sample("S2I_{}".format(t), | ||
| infection_dist(individual_rate=R0 / tau, | ||
| num_susceptible=state["S"], | ||
| num_infectious=state["I"] + X, |
There was a problem hiding this comment.
i guess the X contribution is active at all times?
There was a problem hiding this comment.
Correct. That modeling decision seems more parsimonious, simpler to code, and quantitatively close to the alternative (one initial infection at an unknown time) because during an outbreak, new infections should be dominated by local internal infections.
| :rtype: dict | ||
| """ | ||
| samples = super().predict(forecast) | ||
| samples["first_infection"] = samples["I"].cumsum(-1).eq(0).sum(-1) |
There was a problem hiding this comment.
can you please comment on this quantity?
fehiepsi
reviewed
May 2, 2020
martinjankowiak
previously approved these changes
May 4, 2020
fehiepsi
approved these changes
May 4, 2020
martinjankowiak
approved these changes
May 4, 2020
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Addreses #2426
This adds an
UnknownStartSIRModelthat models infections overpre_window + durationmany time steps. During the initialpre_window-many steps there are zero observed infections but possible "spontaneous" infections from an external source, at a poisson rate of1 / pre_window.To make this model realistic I've also assumed two different detection rates:
rho0up until the first infection andrho1thereafter. This new inhomogeneous functionality requires a newIndex(-)[-]helper similar to our olderVindex(-)[-]. I plan to follow this up with another model with heterogeneous parameters for both ofR0andrho, possibly following brownian motion; this model will make heavier use ofIndex(-)[-].Tested-
UnknownStartSIRModelIndex(-)[-]