Oftentimes, SIR model is used in epidemiology to model infection. Based on the model form, the original SIR model is continuous, and strictly should be solved with continous ode solver. However, it seems to me that the SIR model, provided with day data, originally captures the day dynamics so perhaps using discrete dynamic might be more of the epidemiology case.

So my question is, in epidemiology, which form should be considered as more accurate, or it does not matter at all ?

  • Of the variations on SIR I've heard/read about, I don't recall the discrete one being fashionable.
    – Fizz
    Dec 19 '20 at 8:41
  • But see journals.sagepub.com/doi/pdf/10.1177/0037549716640877 The "event-based" SIR in that paper is a discrete one. The paper only has a handful of citations though, and there's not much real-world validation in it, although it does have some, using the 1918 flu.
    – Fizz
    Dec 19 '20 at 8:48
  • So most of them are solved using ode solver ? I think what confuses me most is the what would be the effect of discrete approximation, e.g., if we want to add into individual factor or group factor as in web.stanford.edu/~chadj/sird-paper.pdf or economics.mit.edu/files/19698, it seems we might have to use discrete version. So how good are they capturing the epidemiology dynamics ?
    – exteral
    Dec 19 '20 at 15:21
  • If you're asking about multi-group models, see this related q: medicalsciences.stackexchange.com/questions/25312/… (not yet answered). I'm not sure there is a "standard"...
    – Fizz
    Dec 19 '20 at 16:25
  • On can analyze multi-group models in the continuous case as well e.g. core.ac.uk/download/pdf/81957441.pdf I'm not sure about the numerical solver details.
    – Fizz
    Dec 19 '20 at 16:32

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.