[R] StructTS and arima and missing values

From: Gavin Simpson <gavin.simpson_at_ucl.ac.uk>
Date: Thu 02 Jun 2005 - 00:14:45 EST

Dear List,

I am thinking about ways in which I might analyse some stratigraphic data. The nature of the data series I have generates a number of issues:

  1. The data I have in mind come from a sediment core sequence taken from the bottom of a lake. The sequence is sliced into a priori defined slices, in this case 0.2cm per slice. in this way a sequence of 0.2cm slices is produced for the entire core.
  2. Each slice is assigned a date (plus some error) using radiometric dating techniques and a derived age/depth model (we age some of the samples and then interpolate/extrapolate for the other samples). This can be done in a variety of ways but effectively the end result is that each 0.2cm sediment slice has a date (year) attached to it (with some error). Changes in the lake system tend to result in changes in the accumulation rate of the sediment sequence, so what we end up with is say a 200 year core sequence that is irregularly sampled in time, but regularly in depth down core.

So for example in one core I end up with the following sequence of years sampled:

> dat

  [1] 2001 2000 1999 1998 1997 1996 1994 1993 1992 1990 1988 1986

[13] 1984 1982 1980 1977 1974 1972 1969 1966 1963 1960 1957 1953
[25] 1950 1946 1943 1940 1936 1931 1927 1922 1918 1914 1908 1902
[37] 1896 1890 1884 1878 1872

I am prepared to accept, for the sake of modelling, that these dates are known and ignore the errors in the dating if that helps.

Having read Brian Ripley's article on Time series in R News Vol 2(2) June 2002, I know that arima and StructTS can now handle missing values, and there is some discussion about the specifics of how these functions can handle missing values, but it is still not clear, in my mind at least, if it would be appropriate to use arima or StructTS on data of this nature -- I'm more interested in fitting a structured time series to this data.

Can StructTS cope with missing values in the sense that I have described them above? If anyone has any suggestions as to how I might approach these data using R they would be gratefully received.

Many thanks for your time,


Gavin Simpson                     [T] +44 (0)20 7679 5522
ENSIS Research Fellow             [F] +44 (0)20 7679 7565
ENSIS Ltd. & ECRC                 [E] gavin.simpsonATNOSPAMucl.ac.uk
UCL Department of Geography       [W] http://www.ucl.ac.uk/~ucfagls/cv/
26 Bedford Way                    [W] http://www.ucl.ac.uk/~ucfagls/
London.  WC1H 0AP.

R-help@stat.math.ethz.ch mailing list
PLEASE do read the posting guide! http://www.R-project.org/posting-guide.html
Received on Thu Jun 02 00:20:11 2005

This archive was generated by hypermail 2.1.8 : Fri 03 Mar 2006 - 03:32:20 EST