R help - Aug 2008 - Time Series w/irregular frequency, how to construct a time series object?

Hello,

I am having trouble constructing a time series object for my data. This is
because the frequency is irregular: one year, there may be only 100
individuals, another 200. There are 100 measurements for every individual. I
have the observations in a data frame with the year that they were taken from
as a factor.

I would like to plot the top 10 mean measurements for each year and connect them
with a line, to measurements in the same category. I would also like to try
using linear filtering and exponential smoothing on the measurements to look
for patterns, in particular synchronized drastic changes in measurement values.

Are all the time-series analysis functions inappropriate because of the
irregular frequency in my data? I looked at the reference manual, and it seems
that even building the tsp attribute manually would require providing a
frequency argument. It's quite possible that I have some very wrong ideas
about
time series analysis, as I am just beginning to study the subject.

Thanks for any help,

--Marco

The zoo package can represent irregular time series.

On Wed, Aug 20, 2008 at 4:41 AM, Marco Leandro Carmosino
<malc07 at hampshire.edu> wrote:
> Hello,
>
> I am having trouble constructing a time series object for my data. This is
> because the frequency is irregular: one year, there may be only 100
> individuals, another 200. There are 100 measurements for every individual.
I
> have the observations in a data frame with the year that they were taken
from
> as a factor.
>
> I would like to plot the top 10 mean measurements for each year and connect
them
> with a line, to measurements in the same category. I would also like to try
> using linear filtering and exponential smoothing on the measurements to
look
> for patterns, in particular synchronized drastic changes in measurement
values.
>
> Are all the time-series analysis functions inappropriate because of the
> irregular frequency in my data? I looked at the reference manual, and it
seems
> that even building the tsp attribute manually would require providing a
> frequency argument. It's quite possible that I have some very wrong
ideas about
> time series analysis, as I am just beginning to study the subject.
>
> Thanks for any help,
>
> --Marco
>
> ______________________________________________
> R-help at r-project.org mailing list
> https://stat.ethz.ch/mailman/listinfo/r-help
> PLEASE do read the posting guide
http://www.R-project.org/posting-guide.html
> and provide commented, minimal, self-contained, reproducible code.
>