Did you know: That Weird Al Yankovic recorded My Balona in a toilet cubical as the reverb was good?Close enoughCovid 19 > Corona virus > The RonaPardon my ignorance, but what is "Rona", is it for grumpiness?
Serving the Quantitative Finance Community
Re: Models for Covid-19
Linkedin: https://www.linkedin.com/in/jasonbelldata/
Author of Machine Learning: Hands on for Developers and Technical Professionals (Wiley).
Contributor: Machine Learning in the City (Wiley).
Author of Machine Learning: Hands on for Developers and Technical Professionals (Wiley).
Contributor: Machine Learning in the City (Wiley).
Re: Models for Covid-19
My model, which I call the minimal SIR model, is a discrete stochastic process that is also a card game.
https://stochanswers.com/education/epidemic.html
Do you know how an epidemic grows? One infects two, two infect four or three or two or one or none. All with different probabilities and a weighted average of logistic growth.
https://stochanswers.com/education/epidemic.html
Do you know how an epidemic grows? One infects two, two infect four or three or two or one or none. All with different probabilities and a weighted average of logistic growth.
-
- katastrofa
- Posts: 7930
- Joined:
- Location: Event Horizon
Re: Models for Covid-19
I think the card game is closer to Galton-Watson branching process, developed by the two to model population growth, rather than to basic SIR. It simplifies it by assuming a constant number of new infections (two).
Re: Models for Covid-19
The game, when scaled, averages to the standard SIR model. If I remember correctly the computer played it with R0 = 2, 3 and 4. The ratio of growth to exponential growth is logit-normally distributed. Note that the model debunks huge numbers of stochastic SIR models in which the modellers just make up any old noise. I also extended the game from complete graphs to social networks showing that lockdowns reduce the number of infected leading to second waves.
-
- katastrofa
- Posts: 7930
- Joined:
- Location: Event Horizon
Re: Models for Covid-19
The noise in stochastic SIR serves to estimate the variability of outcomes - important for evaluating high risk scenarios, designing policies, etc. Averaged trajectories are useless.
Re: Models for Covid-19
No, averaged trajectories serve to validate the model against the standard SIR model. The stochastic model on a complete graph is useless, the average when applied to a social network and its confidence intervals is useful.
-
- katastrofa
- Posts: 7930
- Joined:
- Location: Event Horizon
Re: Models for Covid-19
Yeah, as I said, what matters are variability and confidence bounds, not the average
-
- katastrofa
- Posts: 7930
- Joined:
- Location: Event Horizon
Re: Models for Covid-19
Yeah, as I said, what matters are variability and confidence bounds, not the average