Skip to contents

Create or edit a score specification

Source: R/create_score.R
create_score.Rd

Add to a table of score specifications, where each one is represented by a row. The score specification will only be added if all arguments are valid. Alternatively, edit an existing score specification with the editing parameter.

Usage

create_score(
  scores,
  editing = NA,
  score_type = NULL,
  colname = NULL,
  score_name = NULL,
  weight = NULL,
  lb = NULL,
  ub = NULL,
  centre = NULL,
  inverse = NULL,
  exponential = NULL,
  logarithmic = NULL,
  magnitude = NULL,
  custom_args = NULL
)

Arguments

scores

A data frame of score specifications. See scores_init.

editing

The row number of the score specification to edit. If NA (the default), add the newly created score to the end of scores.

score_type

The method to use when creating the score. Either "Linear", "Peak" or "Custom coordinates".

colname

The name of the column to score.

score_name

The resulting name of the score. If "Default", the score will be given a sensible and informative default name.

weight

The weight of the score when calculating the final score. See score_final().

lb

For a linear or peak score, the lower bound of the score. This will produce a score of 0 in most cases.

ub

For a linear or peak score, the upper bound of the score. This will usually produce a score of 1 for linear scores and a score of 0 for peak scores.

centre

For a peak score, the centre of the score. This will produce a score of 1 unless inverse is TRUE.

inverse

For a peak score, whether to invert the scoring method.

exponential

Whether to apply an exponential transformation to the resulting score.

logarithmic

If exponential is TRUE, whether to invert the exponential transformation.

magnitude

If exponential is TRUE, the numeric magnitude of the exponential transformation.

custom_args

For a custom score, a data frame of coordinates used to create the score.

Value

If any required arguments are not valid, scores is returned as is. If editing is NA, scores is returned with an extra row. If editing is a valid row number, scores with 1 edited row is returned.

Linear scores

When the score_name is "Linear", a linear score is created. To make the score you need to specify the lb and ub arguments.

If the column value is less than or equal to the lb argument, the score is 0. If the column value is more than or equal to the ub argument, the score is 1.

Otherwise, the score is defined is the proportion of the distance of the column value between the lb and ub.

If the lb argument is more than the ub argument, the score is inverted. This means that the lb produces a score of 1, the ub produces a score of 0, etc.

Peak scores

When the score_name is "Peak", a peak score is created. To make the score you need to specify the lb, ub, centre and inverse arguments. The lb, ub and centre arguments must be numeric, and the centre must be between the lb and ub.

If the column value is less than or equal to the lb argument, the score is 0. If the column value is equal to the centre argument, the score is 1. If the column value is more than or equal to the ub argument, the score is 1.

If the column value is in between the lb and centre arguments, the score is defined as the proportion of the column value along between the lb and centre. If the column value is in between the centre and ub arguments, the score is defined as the proportion of the column value along between the ub and centre.

When inverse is TRUE, the score is inverted: the lower bound and upper bound produce a score of 1, and the centre produces a score of 0.

Custom scores

When score_type is "Custom coordinates", a custom score is created. This allows you to define a set of coordinates, where the x coordinate is a value in the column, and the y coordinate is a score between 0 and 1. The score will then be created by connecting the coordinates together. The coordinates should be in the form of a data frame, with the x coordinates in the 'x' column and the y coordinates in the 'y' column.

This can be used to create a huge variety of different scores.

For more detailed information, along with plots that demonstrate the 3 scoring methods, see the Prototype 2 vignette: vignette("prototype2", package = "financialDataAnalysis")

See also

Examples

scores <- create_score(
  scores_init,
  score_type = "Linear", colname = "x", score_name = "Default",
  weight = 1, lb = 1, ub = 6, exponential = FALSE
)

# Change the weight
create_score(
  scores,
  editing = 1, score_type = "Linear", colname = "x",
  score_name = "Default", weight = 2, lb = 1, ub = 6, exponential = FALSE
)
#> # A tibble: 1 × 12
#>   score_type colname score_n…¹ weight    lb    ub centre inverse expon…² logar…³
#>   <chr>      <chr>   <glue>     <dbl> <dbl> <dbl>  <dbl> <lgl>   <lgl>   <lgl>  
#> 1 Linear     x       Score 1:…      2     1     6     NA NA      FALSE   NA     
#> # … with 2 more variables: magnitude <dbl>, custom_args <list>, and abbreviated
#> #   variable names ¹​score_name, ²​exponential, ³​logarithmic