ed: Fit ed model to data
In hafen/ed: ed: a regression approach to density estimation

Description Usage Arguments Details Value Note See Also

View source: R/ed.R

Calculate raw ed density estimates and fit ed model to data using loess

1
2
3

ed(x, k = 10, disjoint = TRUE, degree = 2, span = 0.3, xgrid = 500,
  bounds = c(min(x), max(x)), f = NULL, family = "gaussian",
  normalize = FALSE, delta = 0, lower = NULL, control = loess.control())

`x`	data, a vector
`k`	the gap width (in number of observations) with which to compute the raw estimates
`disjoint`	should non-overlapping gaps be computed? (default `TRUE`)
`degree`	degree of loess fitting (see `loess`)
`span`	span of loess fitting (see `loess`)
`xgrid`	number of equally-spaced points along the support of `x` at which to compute the fit
`bounds`	bounds at which to fit the density (see details)
`f`	a function providing a true density or hypothesized density, with which the ed estimate can be compared (optional)
`family`	loess parameter (`loess`)
`normalize`	should the resulting density be normalized so that it integrates to one?
`delta`	grid augmentation parameter (experimental). A value of 0 (default) disables grid augmentation.
`lower`	grid augmentation parameter (experimental)
`control`	loess parameter (see `loess.control`)

bounds...

a list with a lot of things (to be documented...). For now, look at str(result) to get an idea.

This function is provided as a convenience, but often you may want to simply compute the raw estimates using ed_raw and iteratively figure out how to fit the raw estimates with whatever nonparametric method you like.

ed_raw, ed_plot

hafen/ed documentation built on May 17, 2019, 1:32 p.m.