Modelling of the COVID-19 pandemic: Encouraging forecast of corona infection rates
To predict the future temporal course of the corona pandemic, Prof. Dr. rer. nat. Klaus Hackl, holder of the Chair of Mechanics - Material Theory, has used an epidemic model that is related to formulations from the engineering sciences. According to his prognosis, we can hope for relief in May. In an interview with the PR team of the faculty, he explains in how far his forecast model is different from others.
Prof. Dr. rer. nat. Klaus Hackl used registered infection figures, including data from Germany, Italy and the USA, and showed what development could be expected if current measures were to be maintained. He spoke with the faculty's PR team about the special features of the model, his motivation and his own outlook on the summer semester 2020 at the RUB.
What results does your model rely on and how did you come across them?
"The data used was obtained from online-research. In the process, I came across the known sources (Robert Koch Institute, Johns Hopkins University). I first came up with the model myself (my usual approach in research). As expected, a study of the literature showed that a large variation of similar models already exists. This was my starting point."
What is special about your approach; in how far is it different from other forecasting models?
"I am using a parameter identification based on mathematical optimization for a relatively simple model. The usual approach in epidemiology is different, at least as far as I can judge with my modest knowledge. There, one uses quite complex models and tries to estimate the parameters by means of "expert knowledge", from experience, studies, medical knowledge. However, it's difficult to estimate the reliability of the predictions."
What was your intention?
"When the restrictions began, I wondered, of course, how long they would last. Since there was practically no information about this, I decided to work on it myself. There were already two countries, China and South Korea, with a well advanced process under the restrictions. I wondered to what extent these developments were transferable to our situation. I then decided to share my progress with my colleagues to make the time in the home office a little more interesting."
How resilient do you consider your results to be?
"I am, of course, in some ways an amateur in this field. However, I have tried to obtain reliability of the predictions by using two independent methods of parameter identification, and I have tested this approach for its functionality using data for China and South Korea. The further development so far shows that the forecasts are quite correct."
Based on your results, how do you expect the next two semesters at the RUB and other universities to be organised?
"I assume that the entire summer semester will be digital. It's hard to say what will happen after that. The biggest unsolved problem at the moment is the organisation of examinations with many participants."
A further article on the work of Prof. Dr. rer. nat. Klaus Hackl can be found on the RUB News websites.