class-forecast-multivariate-sample.md
In scoringutils: Utilities for Scoring and Assessing Predictions

as_forecast_multivariate_sample() creates expected structure

Code
  print(result)
Message
  Forecast type: multivariate_sample
  Forecast unit:
  location, model, target_type, target_end_date, and horizon
  Joint across:
  location
Output

            predicted sample_id observed location                 model
                <num>     <int>    <num>   <char>                <char>
      1: 102672.00034         1   106987       DE EuroCOVIDhub-ensemble
      2: 164763.08492         2   106987       DE EuroCOVIDhub-ensemble
      3: 153042.63536         3   106987       DE EuroCOVIDhub-ensemble
      4: 119544.25389         4   106987       DE EuroCOVIDhub-ensemble
      5:  81230.71875         5   106987       DE EuroCOVIDhub-ensemble
     ---                                                               
  35476:    159.84534        36       78       IT  epiforecasts-EpiNow2
  35477:    128.21214        37       78       IT  epiforecasts-EpiNow2
  35478:    190.52560        38       78       IT  epiforecasts-EpiNow2
  35479:    141.06659        39       78       IT  epiforecasts-EpiNow2
  35480:     24.43419        40       78       IT  epiforecasts-EpiNow2
         target_type target_end_date horizon .mv_group_id
              <char>          <Date>   <num>        <int>
      1:       Cases      2021-05-08       1            1
      2:       Cases      2021-05-08       1            1
      3:       Cases      2021-05-08       1            1
      4:       Cases      2021-05-08       1            1
      5:       Cases      2021-05-08       1            1
     ---                                                 
  35476:      Deaths      2021-07-24       2          224
  35477:      Deaths      2021-07-24       2          224
  35478:      Deaths      2021-07-24       2          224
  35479:      Deaths      2021-07-24       2          224
  35480:      Deaths      2021-07-24       2          224

Code
  cat("Class:", class(result), "\n")
Output
  Class: forecast_multivariate_sample forecast data.table data.frame 
Code
  cat("Forecast type:", get_forecast_type(result), "\n")
Output
  Forecast type: multivariate_sample 
Code
  cat("Forecast unit:", toString(get_forecast_unit(result)), "\n")
Output
  Forecast unit: location, model, target_type, target_end_date, horizon 
Code
  cat("Number of rows:", nrow(result), "\n")
Output
  Number of rows: 35480 
Code
  cat("Number of columns:", ncol(result), "\n")
Output
  Number of columns: 9 
Code
  cat("Column names:", toString(names(result)), "\n")
Output
  Column names: predicted, sample_id, observed, location, model, target_type, target_end_date, horizon, .mv_group_id 
Code
  cat("Number of unique groups:", length(unique(result$.mv_group_id)), "\n")
Output
  Number of unique groups: 224

score.forecast_multivariate_sample() creates expected output

Code
  print(scores)
Output
       target_end_date target_type forecast_date                 model horizon
                <Date>      <char>        <Date>                <char>   <num>
    1:      2021-05-08       Cases    2021-05-03 EuroCOVIDhub-ensemble       1
    2:      2021-05-08       Cases    2021-05-03 EuroCOVIDhub-baseline       1
    3:      2021-05-08       Cases    2021-05-03  epiforecasts-EpiNow2       1
    4:      2021-05-08      Deaths    2021-05-03 EuroCOVIDhub-ensemble       1
    5:      2021-05-08      Deaths    2021-05-03 EuroCOVIDhub-baseline       1
   ---                                                                        
  220:      2021-07-24      Deaths    2021-07-12 EuroCOVIDhub-baseline       2
  221:      2021-07-24      Deaths    2021-07-05       UMass-MechBayes       3
  222:      2021-07-24      Deaths    2021-07-12       UMass-MechBayes       2
  223:      2021-07-24      Deaths    2021-07-05  epiforecasts-EpiNow2       3
  224:      2021-07-24      Deaths    2021-07-12  epiforecasts-EpiNow2       2
       energy_score variogram_score .mv_group_id
              <num>           <num>        <int>
    1:   23454.2838     20896.14336           37
    2:   34263.4665     17236.13992           38
    3:   67609.0941    132166.57638           39
    4:     138.9681       210.47403           40
    5:     401.5740      1260.64947           41
   ---                                          
  220:     275.4460       787.91959          256
  221:     142.7219       203.85518          257
  222:      96.7978        72.22572          258
  223:     142.3382       138.22312          259
  224:     137.3695       249.47336          260

Code
  cat("Class:", class(scores), "\n")
Output
  Class: scores data.table data.frame 
Code
  cat("Number of rows:", nrow(scores), "\n")
Output
  Number of rows: 224 
Code
  cat("Number of columns:", ncol(scores), "\n")
Output
  Number of columns: 8 
Code
  cat("Column names:", toString(names(scores)), "\n")
Output
  Column names: target_end_date, target_type, forecast_date, model, horizon, energy_score, variogram_score, .mv_group_id 
Code
  cat("Energy score range:", paste(range(scores$energy_score, na.rm = TRUE),
  collapse = " to "), "\n")
Output
  Energy score range: 37.8373892350605 to 433525.521054322 
Code
  cat("Number of non-NA energy scores:", sum(!is.na(scores$energy_score)), "\n")
Output
  Number of non-NA energy scores: 224 
Code
  cat("Sample of energy scores:", toString(head(scores$energy_score, 5)), "\n")
Output
  Sample of energy scores: 23454.2837522622, 34263.4664789642, 67609.0941230609, 138.968094149042, 401.574009570395

Code
  cat("=== Specific Metrics Test ===\n")
Output
  === Specific Metrics Test ===
Code
  cat("Class:", class(scores_specific), "\n")
Output
  Class: scores data.table data.frame 
Code
  cat("Number of rows:", nrow(scores_specific), "\n")
Output
  Number of rows: 224 
Code
  cat("Number of columns:", ncol(scores_specific), "\n")
Output
  Number of columns: 7 
Code
  cat("Column names:", toString(names(scores_specific)), "\n")
Output
  Column names: target_end_date, target_type, forecast_date, model, horizon, energy_score, .mv_group_id 
Code
  cat("Energy score range:", paste(range(scores_specific$energy_score, na.rm = TRUE),
  collapse = " to "), "\n")
Output
  Energy score range: 37.8373892350605 to 433525.521054322