mk_autoscale_function function

Create a function that transforms a current (multi-dimensional) score into a boxcox normalized one

mk_autoscale_function(
  batch_container,
  scoring,
  random_perm,
  use_boxcox = TRUE,
  sample_attributes_fixed = FALSE
)

Arguments

• batch_container: An instance of BatchContainer .
• scoring: A named list() of scoring function. Each function should return a vector of non-zero length.
• random_perm: Number of random sample permutations for the estimation of the scaling params.
• use_boxcox: Logical; if TRUE and the bestNormalize package is available, boxcox transformations will be used to normalize individual scores. If not possible, scores will just be transformed to a zero mean and unit standard deviation.
• sample_attributes_fixed: Logical; if FALSE, simulate a shuffle function that alters sample attributes at each iteration.

Returns

The transformation function for a new score vector