Range Normalization

Like min_max_norm, but treats every numeric column as coming from a single combined distribution — one global min/max is used to rescale all of them together, rather than normalizing each column independently. Unlike min_max_norm, non-numeric columns are kept (only the numeric ones are transformed) — matching the original implementation this was migrated from.

Usage

range_norm(obj)

Arguments

obj

A numeric vector, matrix, data frame, or tibble.

Value

A normalized object of the same dimensions and class as obj.

Examples

range_norm(data.frame(x = c(1, 2, 3), y = c(-10, 0, 10)))
#>      x   y
#> 1 0.55 0.0
#> 2 0.60 0.5
#> 3 0.65 1.0