kde: Kernel Density Estimate

In BillVenables/kdeR: Kernel Density Estimates

Description

A pure R implementation of a one-dimensional KDE with a gaussian kernel. Mainly of programming interest.

Usage

kde(
  y,
  bw = bw.nrd0(y),
  N = 512,
  lower = ry[1] - 3 * bw,
  upper = ry[2] + 3 * bw
)

## S3 method for class 'kde'
plot(
  x,
  ...,
  col = "steel blue",
  xlab = bquote(x == italic(.(x$data_name))),
  ylab = expression(kde(italic(x)))
)

## S3 method for class 'kde'
print(x, ...)

## S3 method for class 'kde'
as.data.frame(x, row.names, optional, ..., responseName = "density")

Arguments

y	Numeric vector of observed values
bw	Numeric value: bandwidth, standard deviation of the kernel density
N	integer: number of equally spaced points for the kde
lower	Numeric value: lower range limit for the kde
upper	Numeric value: upper range limit for the kde
x	an obeject of class "kde"
...	Extra arguments passed on to methods.
xlab, ylab, col	base graphics parameters
row.names, optional	As for the as.data.frame generic
responseName	character string: name for the response variable

Value

a list object with results of the kde estimation.

Examples

set.seed(1234)
x_value <- rchisq(10000, 2)
kx <- kde(x_value, lower = 0)
kx
if("package:ggplot2" %in% search()) {
  ggplot(as.data.frame(kx)) +
    aes(x = x_value, y = density) +
    geom_line(colour = "steel blue") +
    geom_function(fun = function(x) dchisq(x, 2),
                  colour = "firebrick3") +
    theme_bw() +
    xlab(expression(x == italic(x_value))) +
    ylab(expression(kde(italic(x))))
} else {
  oldPar <- par(mar = c(3, 3, 1, 1) + 0.1, las = 1, mgp = c(1.5, 0.5, 0),
                cex.axis = 0.7, cex.lab = 0.8, tck = -0.015)
  plot(kx, ylim = c(0,0.5))
  curve(dchisq(x, 2), add = TRUE, col = 2, n = 1000)
  rug(x_value, col = "dark green")
  box()
  par(oldPar)
}
rm(oldPar)

BillVenables/kdeR documentation built on April 12, 2021, 1:51 p.m.