Cyrano wrote:Thanks for your posts Amber G.
Is there a summary of the assumptions (or parameters, impact (+/-) and their weightages) of the models you are following? Unless someone understands the assumptions and agrees that are realistic, expressing confidence or lack thereof on any model is usually a gut reaction based to what the model is forecasting.
<snip>
You are welcome.
SUTRA model (and many other recent such models) is quite respectable. As I have posted this link before, one good place is to checkout - for those who have mathematical and scientific background - is arXive
https://arxiv.org/abs/2101.09158. Authors are reputable scientist - Dr. Manindra Agrawal is world renowned
mathematician.
---
-- SUTRA stands for
S - Susceptible (ratio of population which could potentially catch the virus)
U - Undetected ( Causes may be various - not showing symptoms, undercount, authorities not reporting, untested population etc)
T - Tested positive
R - Recovered/Removed (People who died, Recovered, fully vaccinated , or who now have immunity and are no longer in other groups).
The model defines differential equation type relationships using a few basic parameters -- what sets it apart is that it has very good mathematical robustness and powerful statistical ways to determine the basic parameters from time-series data, and other data like sero survey, scientific data about viruses and effect masks etc -- -- even when some quantities (like U - if the testing is not that wide-spread etc) are fuzzy.
Again, am not sure if brf is place so I will stop here.
(I am aware of "criticism" or even abuses heard toward this model in common media , so I am stopping here..sometimes it becomes pointless to "discuss" depending on the audience)
--- Added later: Here is the abstract from the paper I quoted above:
SUTRA: An Approach to Modelling Pandemics with Asymptomatic Patients, and Applications to COVID-19
Manindra Agrawal, Madhuri Kanitkar, Mathukumalli Vidyasagar
In this paper, we present a new mathematical model for pandemics that have asymptomatic patients, called SUTRA. The acronym stands for Susceptible, Undetected, Tested (positive), and Removed Approach. There are several novel features of our proposed model. First, whereas previous papers have divided the patient population into Asymptomatic and Infected, we have explicitly accounted for the fact that, due to contact tracing and other such protocols, some fraction of asymptomatic patients could also be detected; in addition, there would also be large numbers of undetected asymptomatic patients. Second, we have explicitly taken into account the spatial spread of a pandemic over time, through a parameter called "reach." Third, we present numerically stable methods for estimating the parameters in our model.
We have applied our model to predict the progression of the COVID-19 pandemic in several countries. Where data on the number of recovered patients is available, we predict the number of active cases as a function of time. Where recovery data is not available, we predict the number of daily new cases. We present our predictions for three countries with quite distinct types of disease progression, namely: (i) India which has had a smooth rise followed by an equally smooth fall-off in the number of active cases, (ii) Italy, which has witnessed multiple peaks in the number of active cases, and has also witnessed multiple "phases" of the pandemic, and (iii) the USA which has erratic recovery data. In all cases, the predictions closely match the actually observed outcomes.