A data set with binary predictions for COVID-19 cases and deaths constructed from data submitted to the European Forecast Hub.
example_binaryA data frame with 346 rows and 10 columns:
the country for which a prediction was made
name of the country for which a prediction was made
the date for which a prediction was made
the target to be predicted (cases or deaths)
true observed values
the date on which a prediction was made
name of the model that generated the forecasts
forecast horizon in weeks
predicted value
Predictions in the data set were constructed based on the continuous example data by looking at the number of samples below the mean prediction. The outcome was constructed as whether or not the actually observed value was below or above that mean prediction. This should not be understood as sound statistical practice, but rather as a practical way to create an example data set.
The data was created using the script create-example-data.R in the inst/ folder (or the top level folder in a compiled package).