ks (version 1.15.3)

as.kde: Quasi kernel density estimate

Description

Quasi kernel density estimate for gridded data.

Usage

as.kde(x, density)

Value

A quasi-kernel density estimate is an object of class kqde (for density=TRUE) or kqdde (for density=FALSE) which is a list with fields:

x

zero vector or matrix

eval.points

vector or list of points at which the estimate is evaluated - copied from input x

estimate

density estimate - copied from input x

gridded

flag for estimation on a grid

binned

flag for binned estimation

names

variable names

cont

vector of probability contour levels

This mimics the output from kde (for density=TRUE) or kdde(for density=FALSE).

Arguments

x

tidy matrix with first d columns are evaluation points of an estimation grid, and last column is estimate value; or list with elements eval.points, estimate

density

flag to compute bona fide density. Default is TRUE.

Details

This function converts the input x into a quasi-kernel density estimate which is treated similarly to the output from kde or kdde. If density=TRUE then all the estimate values are non-negative. If density=FALSE then the estimate values can be positive or negative. A contourLevels method computes an approximation of the density contour levels, and a plot method computes analogous plots.

See Also

kde, kdde

Examples

## univariate 
ep1 <- seq(0,1, length=401)
est1 <- dbeta(ep1, shape1=2, shape2=5)
f1 <- data.frame(ep1, estimate=est1)
f1.kde <- as.kde(f1)
plot(f1.kde)

## bivariate
ep2 <- expand.grid(x=seq(-pi,pi, length=151), y=seq(-pi,pi, length=151))
est2 <- cos(ep2[,1]*pi/2) + sin(ep2[,2]*pi/2)  
f2 <- data.frame(ep2, estimate=est2)
f2.kde <- as.kde(f2, density=FALSE)
plot(f2.kde, display="filled.contour")
plot(f2.kde, display="persp", phi=10)

## trivariate
mus3 <- c(0,0,0)
Sigmas3 <- 1/2*diag(c(1,1/3,2))
props3 <- 1
ep3 <- expand.grid(x=seq(-3,3, length=51), y=seq(-3,3, length=51), z=seq(-3,3, length=51))
est3 <- dmvnorm.mixt(ep3, mus=mus3, Sigmas=Sigmas3, props=props3)
f3 <- data.frame(ep3, estimate=est3)
f3.kde <- as.kde(f3)
plot(f3.kde)