Skip to main content



Tracking an epidemic requires computer modeling, but what if the models are wrong?

https://phys.org/news/2018-11-tracking-epidemic-requires-modelsbut-wrong.html

 

The article, Tracking an epidemic requires computer modeling, but what if the models are wrong?, discusses the statistical models that attempt to predict the spread of epidemics. The problem of these computer models is the factors that are used in creating them. The first thing is that as the article states: the numbers are never constant. It is important to take into account several components of research before making any conclusions such as where does the person contaminated, live, the places that he goes to on a daily basis and the number of people that their social network is built on. Not only that, but using past results as a form of projection also doesn’t work because of these different factors that are taken into account when projecting a future epidemic.

 

Using statistics from the workplace, from the household, and from friends including strong and weak ties has led to researchers from Northeastern to accomplish an accurate module of how some epidemic spreads in both Netherlands, Italy and other countries that are constituted of networks that share the same attributes. In order to achieve this kind of accuracy for more countries it will be needed to gather much more data in those areas.

 

In class, we have seen the basic model which uses the branching process (Galton-Watson process) designed to calculate the probability of disease spreading. It does that through the use of “branches” that represent the chances of the disease spreading to a certain amount of people and how it exponentially grows until it declines due to the fact that it is unable to spread to people who already caught the disease. Even though it is most likely due to that reason to die out, it is still technically possible to never die out.

 

That model is therefore criticized in the article that clearly states the importance of making a model with specific data for every single area and using one basic process will not be enough to get accurate information in all areas according to the scientists at Northeastern that participated in this research study. However, there are some parts that are matching in both the basic model and what the article discusses. This is the importance of knowing the different factors that will be applied to the data gathered in order to make an accurate model. The deliberate attempts to spread, the obstacles that prevent spreading, the friendships, and the different level of ties are all among the factors that are used in the basic model in order to calculate the probability of a disease spreading and to what extent.

Comments

Leave a Reply

Blogging Calendar

November 2018
M T W T F S S
 1234
567891011
12131415161718
19202122232425
2627282930  

Archives