s-news
[Top] [All Lists]

Re: estimation of the cycle time

To: Chi-tsung Wu <cwu@fcu.edu.tw>
Subject: Re: estimation of the cycle time
From: Prof Brian D Ripley <ripley@stats.ox.ac.uk>
Date: Sun, 23 Sep 2001 07:34:37 +0100 (BST)
Cc: "s-news (E-mail)" <s-news@lists.biostat.wustl.edu>
In-reply-to: <KOEIKNHEGKAAJHHGOIHIKEPMCBAA.cwu@fcu.edu.tw>
On Sun, 23 Sep 2001, Chi-tsung Wu wrote:

> Hi there,
>
> Did someone know a easy or a complicated way to estimate
> the cycle time of a sin curve?

What is observed?  If you observe a sin curve without error in continuous
time, just measure the time between zero-crossings. If the curve is
measured at a grid of times with additive independent errors you want to do
non-linear least-squares fitting (nls, for example).  Very good initial
values can be found by looking at the periodogram.  Complex demodulation
is also relevant.  And so on.

A good overview of the methods available is in

@Book{Bloomfield.00,
  author       = "Bloomfield, P.",
  title        = "Fourier Analysis of Time Series: An Introduction",
  publisher    = "Wiley",
  year         = "2000",
  edition      = "Second",
  ISBN         = "0-471-88948-2",
}

or the 1976 First Edition.


-- 
Brian D. Ripley,                  ripley@stats.ox.ac.uk
Professor of Applied Statistics,  http://www.stats.ox.ac.uk/~ripley/
University of Oxford,             Tel:  +44 1865 272861 (self)
1 South Parks Road,                     +44 1865 272860 (secr)
Oxford OX1 3TG, UK                Fax:  +44 1865 272595


<Prev in Thread] Current Thread [Next in Thread>