% IMPORTANT: The following is UTF-8 encoded. This means that in the presence
% of non-ASCII characters, it will not work with BibTeX 0.99 or older.
% Instead, you should use an up-to-date BibTeX implementation like “bibtex8” or
% “biber”.
@ARTICLE{Sherratt:1009058,
author = {Sherratt, Katharine and Gruson, Hugo and Grah, Rok and
Johnson, Helen and Niehus, Rene and Prasse, Bastian and
Sandmann, Frank and Deuschel, Jannik and Wolffram, Daniel
and Abbott, Sam and Ullrich, Alexander and Gibson, Graham
and Ray, Evan L and Reich, Nicholas G and Sheldon, Daniel
and Wang, Yijin and Wattanachit, Nutcha and Wang, Lijing and
Trnka, Jan and Obozinski, Guillaume and Sun, Tao and Thanou,
Dorina and Pottier, Loic and Krymova, Ekaterina and Meinke,
Jan H and Barbarossa, Maria Vittoria and Leithauser, Neele
and Mohring, Jan and Schneider, Johanna and Wlazlo, Jaroslaw
and Fuhrmann, Jan and Lange, Berit and Rodiah, Isti and
Baccam, Prasith and Gurung, Heidi and Stage, Steven and
Suchoski, Bradley and Budzinski, Jozef and Walraven, Robert
and Villanueva, Inmaculada and Tucek, Vit and Smid, Martin
and Zajicek, Milan and Perez Alvarez, Cesar and Reina, Borja
and Bosse, Nikos I and Meakin, Sophie R and Castro, Lauren
and Fairchild, Geoffrey and Michaud, Isaac and Osthus, Dave
and Alaimo Di Loro, Pierfrancesco and Maruotti, Antonello
and Eclerova, Veronika and Kraus, Andrea and Kraus, David
and Pribylova, Lenka and Dimitris, Bertsimas and Li, Michael
Lingzhi and Saksham, Soni and Dehning, Jonas and Mohr,
Sebastian and Priesemann, Viola and Redlarski, Grzegorz and
Bejar, Benjamin and Ardenghi, Giovanni and Parolini, Nicola
and Ziarelli, Giovanni and Bock, Wolfgang and Heyder, Stefan
and Hotz, Thomas and Singh, David E and Guzman-Merino,
Miguel and Aznarte, Jose L and Morina, David and Alonso,
Sergio and Alvarez, Enric and Lopez, Daniel and Prats, Clara
and Burgard, Jan Pablo and Rodloff, Arne and Zimmermann, Tom
and Kuhlmann, Alexander and Zibert, Janez and Pennoni,
Fulvia and Divino, Fabio and Catala, Marti and Lovison,
Gianfranco and Giudici, Paolo and Tarantino, Barbara and
Bartolucci, Francesco and Jona Lasinio, Giovanna and
Mingione, Marco and Farcomeni, Alessio and Srivastava,
Ajitesh and Montero-Manso, Pablo and Adiga, Aniruddha and
Hurt, Benjamin and Lewis, Bryan and Marathe, Madhav and
Porebski, Przemyslaw and Venkatramanan, Srinivasan and
Bartczuk, Rafal P and Dreger, Filip and Gambin, Anna and
Gogolewski, Krzysztof and Gruziel-Slomka, Magdalena and
Krupa, Bartosz and Moszyński, Antoni and Niedzielewski,
Karol and Nowosielski, Jedrzej and Radwan, Maciej and
Rakowski, Franciszek and Semeniuk, Marcin and Szczurek, Ewa
and Zielinski, Jakub and Kisielewski, Jan and Pabjan,
Barbara and Holger, Kirsten and Kheifetz, Yuri and Scholz,
Markus and Przemyslaw, Biecek and Bodych, Marcin and
Filinski, Maciej and Idzikowski, Radoslaw and Krueger, Tyll
and Ozanski, Tomasz and Bracher, Johannes and Funk,
Sebastian},
title = {{P}redictive performance of multi-model ensemble forecasts
of {COVID}-19 across {E}uropean nations},
journal = {eLife},
volume = {12},
issn = {2050-084X},
address = {Cambridge},
publisher = {eLife Sciences Publications},
reportid = {FZJ-2023-02606},
pages = {e81916},
year = {2023},
abstract = {Background:Short-term forecasts of infectious disease
burden can contribute to situational awareness and aid
capacity planning. Based on best practice in other fields
and recent insights in infectious disease epidemiology, one
can maximise the predictive performance of such forecasts if
multiple models are combined into an ensemble. Here, we
report on the performance of ensembles in predicting
COVID-19 cases and deaths across Europe between 08 March
2021 and 07 March 2022.Methods:We used open-source tools to
develop a public European COVID-19 Forecast Hub. We invited
groups globally to contribute weekly forecasts for COVID-19
cases and deaths reported by a standardised source for 32
countries over the next 1–4 weeks. Teams submitted
forecasts from March 2021 using standardised quantiles of
the predictive distribution. Each week we created an
ensemble forecast, where each predictive quantile was
calculated as the equally-weighted average (initially the
mean and then from 26th July the median) of all individual
models’ predictive quantiles. We measured the performance
of each model using the relative Weighted Interval Score
(WIS), comparing models’ forecast accuracy relative to all
other models. We retrospectively explored alternative
methods for ensemble forecasts, including weighted averages
based on models’ past predictive performance.Results:Over
52 weeks, we collected forecasts from 48 unique models. We
evaluated 29 models’ forecast scores in comparison to the
ensemble model. We found a weekly ensemble had a
consistently strong performance across countries over time.
Across all horizons and locations, the ensemble performed
better on relative WIS than $83\%$ of participating
models’ forecasts of incident cases (with a total N=886
predictions from 23 unique models), and $91\%$ of
participating models’ forecasts of deaths (N=763
predictions from 20 models). Across a 1–4 week time
horizon, ensemble performance declined with longer forecast
periods when forecasting cases, but remained stable over 4
weeks for incident death forecasts. In every forecast across
32 countries, the ensemble outperformed most contributing
models when forecasting either cases or deaths, frequently
outperforming all of its individual component models. Among
several choices of ensemble methods we found that the most
influential and best choice was to use a median average of
models instead of using the mean, regardless of methods of
weighting component forecast models.Conclusions:Our results
support the use of combining forecasts from individual
models into an ensemble in order to improve predictive
performance across epidemiological targets and populations
during infectious disease epidemics. Our findings further
suggest that median ensemble methods yield better predictive
performance more than ones based on means. Our findings also
highlight that forecast consumers should place more weight
on incident death forecasts than incident case forecasts at
forecast horizons greater than 2 weeks.},
cin = {JSC},
ddc = {600},
cid = {I:(DE-Juel1)JSC-20090406},
pnm = {5111 - Domain-Specific Simulation $\&$ Data Life Cycle Labs
(SDLs) and Research Groups (POF4-511)},
pid = {G:(DE-HGF)POF4-5111},
typ = {PUB:(DE-HGF)16},
pubmed = {37083521},
UT = {WOS:001009734700001},
doi = {10.7554/eLife.81916},
url = {https://juser.fz-juelich.de/record/1009058},
}
guest ::