akde: This is a wrapper for the 'akj' adaptive kernel density...
In sethmcg/climod: Bias correction and other tools for climate model output

Description Usage Arguments Details Value Examples

This function calls quantreg::akj and formats its return value as a list with x & y elements, consistent with other kernel density estimation functions. It also provides arguments for defining the x-grid that the PDF is estimated on.

1 2	akde(x, n = 2 * length(x), extend = 3 * stats::bw.nrd(x), min.x = min(x) - extend, max.x = max(x) + extend)

`x`	The data to estimate a PDF for.
`n`	The size of the x-grid to estimate the density on. Defaults to twice the length of x, which is the default used by `akj`.
`extend`	The distance beyond the range of the data to extend the x-grid. Defaults to 3 times the bandwidth calculated by `bw.nrd`, which is close to the default values used by both `density` and `bkde`.
`min.x`	The minimum value of x to compute the density on. Defaults to the minimum data value minus `extend`.
`max.x`	The maximum value of x to compute the density on. Defaults to the maximum data value plus `extend`.

The documentation for akj says that the data values are assumed to be sorted. Although it doesn't appear to matter in practice, this function sorts them just in case.

A list with following components:

x: the x-values where the density is estimated.

y: the estimated density values at each x.

library(KernSmooth)
set.seed(222)
x <- rgamma(300, shape=3, scale=2)
d <- namelist(density, bkde, akde)
p <- lapply(d, do.call, list(x=x))
mplot(p)
legend("topright", names(d), lty=seq(3), col=seq(3))