Transform between output types — convert_output

Transform between output types for each unique combination of task IDs for each model. Conversion must be from a single initial output type to one or more to output types, and the resulting output will only contain the to output types. See details for supported conversions.

Usage

convert_output_type(model_out_tbl, to)

Arguments

model_out_tbl

an object of class model_out_tbl containing predictions with a single, unique value in the output_type column.

to

a named list indicating the desired output types and associated output type IDs. List item name and value pairs may be as follows:

mean: NA (no associated output type ID)
median: NA (no associated output type ID)
quantile: a numeric vector of probability levels OR a dataframe of probability levels and the task ID variables they depend upon. (See examples section for an example of each.) Note that any task ID variable value must appear in the associated model_out_tbl task ID column

Value

object of class model_out_tbl containing (only) predictions of the to output_type(s) for each unique combination of task IDs for each model

Details

Currently, only "sample" can be converted to "mean", "median", or "quantile"

Examples

# We illustrate the conversion between output types using normal distributions
ex_quantiles <- c(0.25, 0.5, 0.75)
model_out_tbl <- expand.grid(
  stringsAsFactors = FALSE,
  group1 = c(1, 2),
  model_id = "A",
  output_type = "sample",
  output_type_id = 1:100
) |>
  dplyr::mutate(value = rnorm(200, mean = group1))

# Output type conversions with vector `to` elements
convert_output_type(model_out_tbl,
  to = list("quantile" = ex_quantiles, "median" = NA)
)
#> # A tibble: 8 × 5
#>   model_id group1 output_type output_type_id value
#> * <chr>     <dbl> <chr>                <dbl> <dbl>
#> 1 A             1 quantile              0.25 0.375
#> 2 A             1 quantile              0.5  1.09 
#> 3 A             1 quantile              0.75 1.75 
#> 4 A             2 quantile              0.25 1.61 
#> 5 A             2 quantile              0.5  2.13 
#> 6 A             2 quantile              0.75 2.81 
#> 7 A             1 median               NA    1.09 
#> 8 A             2 median               NA    2.13 

# Output type conversion with dataframe `to` element
# Output type ID values (quantile levels) are determined by group1 value
quantile_levels <- rbind(
  data.frame(group1 = 1, output_type_id = 0.5),
  data.frame(group1 = 2, output_type_id = c(0.25, 0.5, 0.75))
)
convert_output_type(model_out_tbl,
  to = list("quantile" = quantile_levels)
)
#> # A tibble: 4 × 5
#>   model_id group1 output_type output_type_id value
#> * <chr>     <dbl> <chr>                <dbl> <dbl>
#> 1 A             1 quantile              0.5   1.09
#> 2 A             2 quantile              0.25  1.61
#> 3 A             2 quantile              0.5   2.13
#> 4 A             2 quantile              0.75  2.81