|
|
Curve Fitting the Coronavirus
|
The Coronavirus (2019 Novel Coronavirus) has been at the center of attention for the World Health Organization, and rightfully so.
It is a highly contagious and difficult to treat disease that has now affected ten thousand individuals.
One tool that can assist in projecting the growth of the current outbreak is a tool called curve fitting.
Curve fitting takes historical data (it could be your business data, or in this case medical data on the Coronavirus), and finds the mathematical equation that best fits the historical data.
This mathematical equation is then utilized to project how the data points will continue into the future.
Even though we only have 12 days worth of data, we are able to create projections, and more importantly, gauge the accuracy of those projections using something called the Coefficient of Determination, or R squared.
The closer the R squared value is to 1, the more accurate the equation fits the existing data, and thus the more accurate the projection.
As can be seen from the graphs below, utilizing data obtained from the WHO combined with curve fitting, the next week is really crucial in this outbreak.
The projections which have the best R squared values show that the Coronavirus spread may peak in about a week.
For example, the graph above has an R squared value of .9993, which is pretty accurate.
This graph projects that the outbreak should peak in about a week and be completely contained in about two weeks.
However, ever so slight variations of these models show that the outbreak, based on its current status, could very well spread to every country in the world and spread to over 1 million individuals.
In the graph above, the best estimate, with an R squared value of .9993, is compared with the next best estimate, which has an R squared value of .9986 (a difference of 1/10 of 1%).
This less accurate estimate, shown above by itself, predicts the current outbreak spreading into a world wide epidemic infecting about a million people within about two months.
Based on the current rates, the Coronavirus has been primarily limited to China (99% of the cases are in China).
The most accurate projections show that these cases outside of China should stay below 200 (an R squared value of .9999-an almost perfect fit).
However, the next less accurate model, with an R squared value of .9996, shows that the Coronavirus outbreak could easily grow to thousands of cases outside China within a month.
As WHO mentions, countries without significant resources to contain coronavirus appear to be the largest risk at this time.
They could determine whether the issue subsides in a week or two, affecting only another 20 countries (40 total) and limiting the total cases to 20,000, or whether it spreads to every country and infects millions.
Let's hope the most accurate models are correct, that the efforts to quarantine those affected are successful, and that the entire episode is contained within two weeks.
Does your business ever feel awash in data, unable to gain insights into what your corporate data means?
Could your business benefit from better vision into your corporate data?
Contact NPM Technologies today to benefit from the latest business intelligence methods to facilitate your company's success. 855-NPM-TECH or sales@npmit.com or www.npmit.com.
#coronavirus #novelcoronavirus #2019novelcoronavirus #curvefitting #npm #npmtechnologies #who #worldhealthorganization #cdc #centerfordiseasecontrol #quarantine #businessintelligence #bi
|
|
|