AI-Based Algorithm from Wits Tracks Chances of a COVID-19 Third Wave in SA

6 May

As a service provider of online cloud backup, we often focus our attention on topics related to data. In this article, we would like to give a shoutout to a team that has developed an early detection system that helps to predict future cases of COVID-19.

Enter the spotlight; our very own Wits University as a frontrunner for developing an Artificial Intelligence-based (AI-based) algorithm that has helped to predict the severity/chances of a third wave of COVID-19 South Africa. While the statistics predict that the chances are low, the country is still vulnerable and it’s important that we all continue to do our bit to follow all the necessary COVID protocol and regulations.

The system was developed by Wits University in partnership with iThemba LABS (a leading research facility for accelerator-based science), the Gauteng Provincial Government, and Canada’s York University.

We take a look into the AI-based algorithm and exactly what it seeks to achieve.

what is the functionality of the new AI-based algorithm?

The AI-Based algorithm functions as an early detection warning system. It predicts daily cases in the future based on historical data from past infection rates across the country. While the algorithm works independently, there is another algorithm that is also referenced. The existing algorithm currently used by the SA government is based on classic analytics, and when run at the same time as the Wits University algorithm, more reliability is added to the process and outcomes.

Each algorithm works independently, navigating a large number of parameters, dimensions and analysing complex problems while continuously learning from the data.

what’s the benefit of another detection system?

The general public and the health care industry need to know beyond reasonable doubt that they can rely on the systems in place. With two detection systems running simultaneously and updated daily, the results are more reliable. Should the systems show a discrepancy, it provides a red flag worth further investigation.

The rise and fall of COVID-19 is still very much an unknown entity but it’s evident that infection waves are hard to predict or control. An AI-based algorithm provides a warning signal to policymakers and the population in South Africa.

The system monitors daily data, and when it notices a change that may indicate a new wave or greater threat, it issues an alert to the relevant decision-makers and authorities.

are AI algorithms accurate?

It needs to be noted that predictions by algorithms can be accurate, but there is never 100% guaranteed accuracy. It is better to use the system to predict outcomes over the short term (a few weeks at a time) than over the long term (say, a year), as the longer the prediction period, the less accurate the predictions stand to be.

It’s undeniable that such systems provide a valuable service, especially when quick and decisive action needs to be taken

last word

Thanks to the enthusiasm and commitment of local students at the University you can follow these predictions for each of our provinces on the 3rd wave risk index dashboard. The orange line seen in the graphs indicates a possible risk for an onset of a new wave in the respective province.

A big shout out to Wits University and all their co-collaborators! Well done!

Comments are closed.