Skip to contents

Scaling weightings

Source: R/helpers.R
scale_weights.Rd

Scales quantitative errors or qualities of orientation measurements. Useful for some statistical summaries or fault-slip inversion.

Usage

scale_weights(
  e,
  error_type = c("rank", "angle", "rup"),
  scaling = c("lin", "inv_lin", "inv_square", "exp"),
  replace_na = TRUE,
  norm = FALSE
)

Arguments

e

numeric. Weights

error_type

character. One of "rank" (a numeric value ranking measurement quality), "angle" (a reading error expressed as angles in degrees), and "rup" (RUP values from a previous fault-slip inversion)

scaling

character. Scaling function to use. One of "lin" (leaves weights \(e\) as is), "inv_lin" (\(1/e\)), "inv_square" (\(1/e^2\)), and "exp" (\(\exp{-(x-1)}\))

replace_na

logical. Imputation? Whether NA should be replaced by the mean of the weights? (TRUE by default)

norm

logical. Whether the scaled weights should be normalized by their mean? (FALSE by default)

Value

numeric.

See also

slip_inversion_angelier()

Examples

set.seed(20250411)
# Generate some random weights from 1 (poor) to 5 (good)
err <- sample(1:5, size = 10, replace = TRUE)

# Introduce 3 random NAs
err[sample(length(err), 3)] <- NA

scale_weights(err, error_type = 'rank', scaling = 'inv_square', norm = TRUE)
#>  [1] 0.8095728 0.8095728 0.8095728 0.4817784 3.0111151 0.7527788 0.7527788
#>  [8] 0.4817784 1.3382734 0.7527788