R help - Oct 2002 - s.window in stl()

Hi,

This is actually a theory question.

I'm a bit confused by the s.window parameter in the stl() function (which
is in the ts package).  For example, in the stl documentation it uses the
nottem data, and then:
  plot(stl(nottem, s.win = 4, t.win = 50, t.jump = 1))

What does it mean by s.win = 4?  Is it because a year has 4 seasons
(namely Spring, Summer, Autumn and Winter)?  If so will it make sense if I
use 12 for the 12 months (as I've got a similar data set, and wondering
about this s.window when I fit the data with stl()).

Cheers,

Kevin

------------------------------------------------------------------------------
Ko-Kang Kevin Wang
Postgraduate PGDipSci Student
Department of Statistics
University of Auckland
New Zealand
Homepage: http://www.stat.auckland.ac.nz/~kwan022


-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-
r-help mailing list -- Read http://www.ci.tuwien.ac.at/~hornik/R/R-FAQ.html
Send "info", "help", or "[un]subscribe"
(in the "body", not the subject !)  To: r-help-request at
stat.math.ethz.ch
_._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._

Hi,

I think s.window is the span is lags of loess window. 
Loess centers a window for each time and fits a line by weighted 
least-squares (more weight is given to observations close to the middle of 
the window); the residual (vertical) distance from the line to each point in 
the window is determined. The outliers (points with large residuals) are then 
downweighted and the line is re-estimated; this process is iterated a few 
times. 

So with monthly data you don't have to use s.window=12 (and notice it must
be
odd!)

Hope it helps


Susana
-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-
r-help mailing list -- Read http://www.ci.tuwien.ac.at/~hornik/R/R-FAQ.html
Send "info", "help", or "[un]subscribe"
(in the "body", not the subject !)  To: r-help-request at
stat.math.ethz.ch
_._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._