gsmoothr (version 0.1.7)

tmeanC: Trimmed Mean Smoother

Description

A fast trimmed mean smoother (using C code) of data at discrete points (e.g. probe-level data).

Usage

tmeanC(sp, x, spout = NULL, nProbes = 10, probeWindow = 600, trim = 0.1)

Arguments

sp

numeric vector of positions (x-values)

x

numeric vector of data (corresponding to sp

spout

optional vector of output values to calculate trimmed mean at, default: NULL

nProbes

minimum number of observations required within window

probeWindow

distance (in x) in each direction to look for observations to be used in the trimmed mean

trim

proportion of trim to use in trimmed mean

Value

vector (of the same length as sp (or spout)) giving the trimmed mean smoothed values

Details

Using the specified probe window, this procedure uses all values within the window and calculates a trimmed mean with the specified amount of trim. If there are not enough observations within the window at a given position (as given by nProbes), a zero is returned.

See Also

trimmedMean

Examples

sp <- seq(100, 1000, by=100)
ss <- seq(100,1000, by=50)
set.seed(14)
x <- rnorm(length(sp))

tmC <- tmeanC(sp, x, probeWindow=300, nProbes=5)
tmC1 <- tmeanC(sp, x, spout=sp, probeWindow=300, nProbes=5)
tmC2 <- tmeanC(sp, x, spout=ss, probeWindow=300, nProbes=5)

cbind(tmC,tmC1)

plot(sp, x, type="h", ylim=c(-2,2))
lines(sp, tmC1, col="blue")
lines(ss, tmC2, col="red")