[R-SIG-Finance] Inhomogeneous Time Series Operator

[Jeffrey Ryan]

A few years back I wrote an iEMA.kernel  function in R to attempt to
get the inhomgenous stuff into R.  Josh (TTR) and I discussed adding
it in, and had chatted with Eric Zivot about it (2009??)  -- since, if
I recall, he authored the original S+ version.

My issue was a lack of data to compare output.  I think I was trying
to use the limited output from Eric's book and the user manual - and
that wasn't really the best way to go.

If someone has the interest to test output from R functions against
the known output from S+Finmetrics or some other software
implementation, I'd could likely dig up the code I wrote.

I'd likely aim to add this into TTR or some offshoot package.  There
are some other issues that may make this a non-starter, but anyone
interested can certainly contact me.

Best,
Jeff

On Thu, Feb 10, 2011 at 7:58 AM, Brian G. Peterson <brian at braverock.com> wrote:
> On 02/10/2011 12:42 AM, Jiwon Kim wrote:
>>
>> Hello everyone,
>>
>> I'm wondering if operators for irregulary-spaced time series are
>> available.
>> Examples are those inhomogenous operators introduced in "Operators on
>> *inhomogeneous
>> time
>> series*<http://www.bilokon.co.uk/archive/academic/papers/finance/Zumbach2000.pdf>"
>> by G. Zumbach et. al.
>> It seems that all of those operators are included in S+FinMetrics for
>> S-Plus, but cannot find any resources for R.
>> I was thinking of reading S+FinMetrics implementation, but I don't have an
>> access to S-Plus so don't know how I should proceed.
>> If no package is yet available, it would be nice to get some advice from
>> you
>> experts where I should start.
>
> It seems from a quick glance at the paper that all the metrics are
> calculated not directly over irregular time series but rather some other
> period is chosen to make them regular so that more traditional methods may
> be applied.
>
> xts has very good support for high frequency irregular series.
>
> period.apply and to.period should allow you to homogenize (regularize) the
> series in any way that is appropriate to your needs.
>
> Most of the technical indicators (including EMA, used heavily by the paper)
> are in the TTR package.
>
> The kernel estimators look pretty simple, and proceed from the above. there
> are also many kernel functions available in various R packages. I didn't
> read the paper closely enough, nor did I search on places like r-seek or
> RSiteSearch , for the specific kernel estimator used by the authors of the
> paper.
>
> Regards,
>
>   - Brian
>
> --
> Brian G. Peterson
> http://braverock.com/brian/
> Ph: 773-459-4973
> IM: bgpbraverock
>
> _______________________________________________
> R-SIG-Finance at r-project.org mailing list
> https://stat.ethz.ch/mailman/listinfo/r-sig-finance
> -- Subscriber-posting only. If you want to post, subscribe first.
> -- Also note that this is not the r-help list where general R questions
> should go.
>



-- 
Jeffrey Ryan
jeffrey.ryan at lemnica.com

www.lemnica.com