[Zurück]

N. Ukaj, St. Scheiner, Ch. Hellmich:
"Toward "hereditary epidemiology": A temporal Boltzmann approach to COVID-19 fatality trends";
Applied Physics Reviews, 8 (2021), 4; 24 S.

Countless research contributions reflect two major concepts for modeling the spread of the COVID-19 pandemic: (i) ordinary differential equations for population compartments, such as infected or deceased persons (these approaches often exhibit limited predictive capabilities); and (ii) rules applied to digitally realized agents in the populations (these approaches often lack reliable input data and may become computationally overly expensive). As a remedy, we here introduce and discuss convolutional integrodifferential equations adapted from
Boltzmann´s hereditary mechanics, so as to predict COVID-19 fatality trends from the evolutions of newly infected persons. Replacing the
classical statistical reasoning by deliberations arising from the notion of "virus loads" and the corresponding compliance of the infected population to these loads, model errors with respect to data recorded in 102 countries, territories, or US states can be drastically reduced, namely, up to 98% when compared to the traditional kinetics equation of Kermack and McKendrick. The coefficients of determination between model predictions and recorded data range from 94% to 100%, a precision hitherto unachieved in equation-based epidemic modeling.

COVID-19 pandemic; coefficient of determination; model prediction

http://dx.doi.org/10.1063/5.0062867

https://publik.tuwien.ac.at/files/publik_302995.pdf