Unit:
Κατεύθυνση ΒιοστατιστικήLibrary of the School of Health Sciences
Supervisors info:
Δημήτριος Καρλής, Καθηγητής, Τμήμα Στατιστικής, Οικονομικό Πανεπιστήμιο Αθηνών
Απόστολος Μπουρνέτας, Καθηγητής, Τμήμα Μαθηματικών, Εθνικό και Καποδιστριακό Πανεπιστήμιο Αθηνών
Νικόλαος Δεμίρης, Επικ. Καθηγητής, Τμήμα Στατιστικής, Οικονομικό Πανεπιστήμιο Αθηνών
Original Title:
Spatiotemporal clustering with an application on COVID-19 deaths in the provinces of the Netherlands
Translated title:
Spatiotemporal clustering with an application on COVID-19 deaths in the provinces of the Netherlands
Summary:
Spatio-temporal models for count data are an important tool for a wide range of scientific fields. Recently, they have become particularly crucial since they can be employed to monitor the contagion dynamics of the COVID-19 pandemic, both in time and in space. Considering the endemic-epidemic framework, we first describe the general modelling approach and then employ various extensions. The models are exemplified through an analysis of daily COVID-19 death counts from the twelve provinces of The Netherlands during the first eight months of 2021. Since similar spatial behavior is a common feature of discrete-valued time series data, it needs to be taken into account appropriately. In this paper, we propose the incorporation of an algorithm that will cluster regions based on their spatio-temporal characteristics. In our application, we find that the region specific extensions of the endemic-epidemic model provide a better fit. However, notably, the performance of all the extensions is considerably improved by the incorporation of the clustering algorithm.
Main subject category:
Health Sciences
Keywords:
Clustering, Spatiotemporal model, Endemic-Epidemic model, Netherlands, Provinces, Count data, COVID-19, Coronavirus, Extensions, EM-algorithm, Finite mixture models
