In this page, I will try to summarize Covid 19 modeling.
Many people have been modeling diseases specifically infectious disease and how they can become pandemic. In modeling and forecasting hindsight is 20/20. The modeling can be as good as the data can be.
In the world of AI and ML, there are two paths to modeling. Once our good old mathematical model that represent a known set of data and then you forecast from it. Second one also know as Neural Networks that basically automatically creates a memory of data so that the forecast can be made accurately.
Infectious Disease Modelers have been using SEIRS models for a while with a lot of success rate. Having said that in disease modeling, it is better to be conservative and wrong than right. You are hoping that decision makers will see you model forecast and realize that action must be taken for public safety and thus the impact is much less.
SEIR and SEIRS models
Tried and Truth
References:
- https://en.wikipedia.org/wiki/Mathematical_modelling_of_infectious_disease
- https://www.idmod.org/docs/emod/hiv/model-seir.html
- https://github.com/covid-projections/covid-data-model — covid data models
- https://www.thelancet.com/coronavirus – covid 19 resource center
- https://covidactnow.org/?s=936187 – a page with Covid visualization
- https://gabgoh.github.io/COVID/index.html – calculator so show how things look like
- https://github.com/google-research/open-covid-19-data – open data set
- https://github.com/nytimes/covid-19-data/tree/master/live – New York time’s data set