Every day for the past few weeks, charts and graphs plotting the projected apex of COVID-19 infections have been circulated across newspapers and cable news. Many of these models have been constructed using data from studies on earlier outbreaks like SARS or MERS.
Now, a crew of engineers at MIT has built a model that uses data from the COVID-19 pandemic along with a neural network to find out the efficacy of quarantine measures and better predict the spread of the coronavirus.
Most models used to foretell the spread of a disease observe what is known as the SEIR model, which groups people into susceptible, uncovered, contaminated, and recovered.
Dandekar and Barbastathis revamped the SEIR model by training a neural network to seize the number of infected people who’re under quarantine, and therefore not spreading the infection to others.
The model finds that in locations like South Korea, where there was quick government intervention in implementing robust quarantine measures, the virus spread plateaued more rapidly.
In places that were slower to implement government interventions, like Italy and the U.S., the efficient reproduction number of COVID-19 stays better than one, which means the virus has continued to spread exponentially.
The machine learning algorithm reveals that with the current quarantine measures in place, the plateau for both Italy and the USA will arrive somewhere between April 15-20.