Time series decomposition using a running median
This is a nonlinear filtering method used to smooth, but also to segment a time series. The isolated peaks and pits are leveraged by this method.
decmedian(x, type="additive", order=1, times=1, ends="fill")
x |
a regular time series ('rts' under S+ and 'ts' under R) |
type |
the type of model, either |
order |
the window used for the running median corresponds to 2*order + 1 |
times |
the number of times the running median is applied. By default, 1 |
ends |
the method used to calculate ends. Either "NAs" (fill extremes, non-calculable values with NAs), or "fill" (fill these extremes with the closest calculable median) |
a 'tsd' object
Frédéric Ibanez (ibanez@obs-vlfr.fr), Philippe Grosjean (phgrosjean@sciviews.org)
Gebski, V.J., 1985. Some properties of splicing when applied to non-linear smoothers. Comp. Stat. Data Anal., 3:151-157.
Philips, L. & R. Blomme, 1973. Analyse chronologique. Université Catholique de Louvain. Vander ed. 339 pp.
Tukey, J.W., 1977. Exploratory Data Analysis. Reading Massachusetts: Addison-Wesley.
data(marbio)
ClausoB.ts <- ts(log(marbio$ClausocalanusB + 1))
ClausoB.dec <- decmedian(ClausoB.ts, order=2, times=10, ends="fill")
plot(ClausoB.dec, col=c(1, 4, 2), xlab="stations")
# This is a transect across a frontal zone:
plot(ClausoB.dec, col=c(0, 2), xlab="stations", stack=FALSE, resid=FALSE)
lines(c(17, 17), c(0, 10), col=4, lty=2)
lines(c(25, 25), c(0, 10), col=4, lty=2)
lines(c(30, 30), c(0, 10), col=4, lty=2)
lines(c(41, 41), c(0, 10), col=4, lty=2)
lines(c(46, 46), c(0, 10), col=4, lty=2)
text(c(8.5, 21, 27.5, 35, 43.5, 57), 8.7, labels=c("Peripheral Zone", "D1",
"C", "Front", "D2", "Central Zone"))Please choose more modern alternatives, such as Google Chrome or Mozilla Firefox.