We gratefully acknowledge support from
the Simons Foundation and member institutions.
Full-text links:


Current browse context:


Change to browse by:

References & Citations


(what is this?)
CiteULike logo BibSonomy logo Mendeley logo del.icio.us logo Digg logo Reddit logo ScienceWISE logo

Statistics > Methodology

Title: Cluster-based dual evolution for multivariate time series: analyzing COVID-19

Abstract: This paper proposes a cluster-based method to analyze the evolution of multivariate time series and applies this to the COVID-19 pandemic. On each day, we partition countries into clusters according to both their case and death counts. The total number of clusters and individual countries' cluster memberships are algorithmically determined. We study the change in both quantities over time, demonstrating a close similarity in the evolution of cases and deaths. The changing number of clusters of the case counts precedes that of the death counts by 32 days. On the other hand, there is an optimal offset of 16 days with respect to the greatest consistency between cluster groupings, determined by a new method of comparing affinity matrices. With this offset in mind, we identify anomalous countries in the progression from COVID-19 cases to deaths. This analysis can aid in highlighting the most and least significant public policies in minimizing a country's COVID-19 mortality rate.
Comments: Accepted manuscript. Major expression edits and minor new inferences relative to v2. Equal contribution
Subjects: Methodology (stat.ME); Dynamical Systems (math.DS); Physics and Society (physics.soc-ph); Populations and Evolution (q-bio.PE)
Journal reference: Chaos 30, 061108 (2020)
DOI: 10.1063/5.0013156
Cite as: arXiv:2005.02125 [stat.ME]
  (or arXiv:2005.02125v3 [stat.ME] for this version)

Submission history

From: Nicholas James [view email]
[v1] Tue, 5 May 2020 13:16:54 GMT (1574kb,D)
[v2] Fri, 8 May 2020 16:17:28 GMT (717kb,D)
[v3] Sun, 14 Jun 2020 08:38:58 GMT (729kb,D)

Link back to: arXiv, form interface, contact.