No Thumbnail Available

The SIRD model: A Time-dependent parameter approach

Doraghi, Mehrdaad
(2022)

Files

Doraghi_55141800_2022.pdf
  • Closed access
  • Adobe PDF
  • 781.92 KB

Details

Supervisors
Hafner, Christian
Degree label
Master [120] en statistique, orientation générale, à finalité approfondie
Abstract
Many external elements had played a role in the COVID-19 pandemic. Changes in the population's behaviour, political measures or virus mutations are only few examples. The SIR-like models (Kermack and McKendrick (1927)) are oversimplified epidemiological models that assume a constant transmission rate or constant death rate across time. This master's thesis develops two models to relax these cited assumptions: the semi-parametric SIRD and the Fourier Flexible Form SIRD (FFF SIRD). The latter proves that the transmission rate does vary across time. Correcting the FFF SIRD model for group heteroskedasticity in the data gives more efficient estimates. After selecting a dataset with a two-wave infectious curve, the corrected FFF SIRD model found the expected two-wave result.