team_name: AMM
model_name: EpiInvert
model_abbr: AMM-EpiInvert
model_contributors: 
 - name: Luis Alvarez
   affiliation: Universidad de Las Palmas de Gran Canaria, Spain
   email: lalvarez@ulpgc.es
 - name: Jean-David Morel
   affiliation: Ecole Polytechnique Fédérale de Lausanne, Switzerland
   email: jeandavid.morel@gmail.com
 - name: Jean-Michel Morel
   affiliation: Université Paris-Saclay, France
   email: moreljeanmichel@gmail.com
website_url: https://github.com/lalvarezmat/EpiInvert
license: cc-by-4.0
team_model_designation: primary
methods: Learning from the past a short time forecast of the COVID-19 incidence curve trend.
methods_long: We design a learning procedure, EpiInvertForecast,  to forecast the future trend of the daily incidence. 
  This learning procedure is based on a database of 27,418 incidence trend curves computed in the past, with EpiInvert, 
  using real data. The forecast of the current incidence trend is obtained as a weighted average of the 27,418 database 
  curves where the weight of each database curve depends on the similarity between the current curve and the database 
  curve in the past. We also compute empirical confidence intervals for the forecast estimation and, using the weekly
  seasonality computed by EpiInvert, we also estimate a forecast of the original incidence curve. 
citation: https://doi.org/10.1073/pnas.2105112118 , https://doi.org/10.3390/biology11040540 , 
  https://ctim.ulpgc.es/covid19/EpiInvertForecast.html
data_inputs: https://github.com/covid19-forecast-hub-europe/covid19-forecast-hub-europe/tree/main/data-truth/JHU