Skip to content

A QUESTION ABOUT TIME SERIES

6 messages · Castro, Marcelo T G (GEL, MSX), Alexandre Fayolle, Uwe Ligges +2 more

Original
Castro, Marcelo T G (GEL, MSX)
# 
Dear sirs,

My name is Marcelo Grimaldi and i work for General Electric in the Six Sigma
Program. Recently I became very interested in spectral analysis, because I
think it can be applied to my job. I use Minitab 12 and i have heard that
writing macros for spectral analysis in this software is very difficult.
Could I perform this analysis in a easier way throught "R" ? Have you ever
heard about anybody that perform it using "R"? I would apreciatte very much
if I could get in touch with these specialists.

THANK YOU VERY MUCH,
Marcelo Grimaldi
Six Sigma Black Belt - Manufacturing
GE Lighting Brazil
  

-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-
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
_._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._
Ben Bolker
# 
Yes, through some built-in code and through an additional package.

  I don't know about "getting in touch with the specialists", though --
your best bet is to start with a good general-purpose time series analysis
book, which will have something about spectral analysis and will lead you
to more specialized references.  Or start from whatever references
suggested spectral analysis to you in the first place.  Or follow
interesting references in the documentation of the R functions.

  If you have a working knowledge of R (from reading documentation and the
various public domain guides accessible through CRAN, or the books listed
at CRAN) and a reasonable knowledge of spectral analysis it should be
quite straightforward to use the SA tools in R.

  It's best to save R-help for specific problems with R, although the
occasional interesting statistical problem will probably get responses ...

  [If I were a better Perl programmer I would write a script that would
try to answer "Can R ... ?" questions automatically by (1) searching the
help pages on CRAN, (2) searching the package sources, (3) searching the
mail archives ...]
On Mon, 15 Jan 2001, Castro, Marcelo T G (GEL, MSX) wrote:

            

      
      
      
    

        

  
    

      
      
    

  


  


      

    

    

      
      

        


  

        

      Alexandre Fayolle
    

    

      
      #
    

  


  

      
On Mon, 15 Jan 2001, Castro, Marcelo T G (GEL, MSX) wrote:

        

            

      
      
      
    

        
I don't know about spectral analysis within R, but I am pretty sure that
such thing is implemented in Matlab and Scilab (which is a free clone of
Matlab by INRIA), and maybe in Octave (yet another free clone of Matlab,
part of the GNU project).

Alexandre Fayolle

  
    

      
      
    

  


  


      

    

    

      
      

        


  

        

      Uwe Ligges
    

    

      
      #
    

  


  

      
On Mon, 15 Jan 2001, Alexandre Fayolle wrote:

        

            

      
      
      
    

        
Now I cannot resist to answer to the list (after having sent a private
answer) in the spirit of Ben Bolker's answer:


The real question is: What kind of spectral analysis?

If you want to have a look at the spectrum, the periodogram is the first
thing I am thinking about (an inconsistant(!) estimator for the spectral
density). You can easily calculate it via FFT. Or use
e.g. spectrum(), a user-friendly function which calculates the logarithmic
periodogram [details in the help pages ...].

Uwe Ligges

-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-
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
_._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._

  
  


  


      

    

    

      
      

        


  

        

      Jason Turner
    

    

      
      #
    

  


  

      
On Mon, Jan 15, 2001 at 01:12:36PM -0500, ben at zoo.ufl.edu wrote:

            

      
      
      
    

        
It's not quite the answer you're looking for, but...

I run R on my Linux box, which also runs Apache (www.apache.org).
I've set up the html documentation in the publicly available
docs (visible to apache), and set up htdig as my search engine
(www.htdig.org).  I just type in the subject, and get links to
the (very good) html documentation.  This has served my purposes
*very* well, and I'd recommend it any day.  htdig can be quite
flexible about where it searches, and you could re-direct it
to the CRAN site.  I'm on a slow modem, so I haven't done this.

I do this with any html documentation I get, and it's a good
way to find other software that does the job you want in software
you mightn't have thought of.

Hope that helps

Jason

  
    

      
      
    

  


  


      

    

    

      
      

        


  

        

      Ben Bolker
    

    

      
      #
    

  


  

      
Interesting, and certainly a better solution than hacking Perl scripts.
I was actually more interested in this scripting/htdig option as a
function for other people (e.g., an option set up on the R project web
page) than for myself.  For myself, I'm generally happy enough searching
my local HTML help and searching the packages page on CRAN.  It just seems
that there are many people who are a little too ignorant (no pejorative
connotation intended) or lazy (slight p.c.) to do this search for
themselves, and it might help to have such an engine available.
   Speaking of lazy, the odds of my setting this up are slim, but it might
happen ...

  Ben Bolker

        
On Tue, 16 Jan 2001, Jason Turner wrote: