The signed logarithm allows a pseudo-logarithmic re-scaling to be performed on data that is both positive and negativeis. It defined as:

$$ \textrm{signed_log}_b(x) = \textrm{sign}(x)\log_b(|x| + 1)$$

signed_log(x, base = exp(1))

unsigned_log(x, base = exp(1))

Arguments

x

a numeric vector.

base

a positive number

Examples

x <- seq(-10, 10, 0.01) y <- signed_log(x) plot(x, y, type = "l")
x0 <- unsigned_log(y) plot(y, x0, type = "l")
plot(x, x0, type = "l")