[R-SIG-Finance] PerformanceAnalytics CoSkewness Function

[Brian G. Peterson]

On 01/18/2013 12:58 AM, jkaprich at gmail.com wrote:
> Running ES using...
>
> Method = "modified"
>
> Data is a timeSeries object (SC.Adjusted) with 12 variables.
>
> When calculating m3, do I follow ...
>
> CoSkewness (SC.Adjusted, SC.Adjusted)
>
> Appreciate your help.

What about the documentation isn't clear?

You haven't followed the posting guide and provided a reproducible 
example, so it is rather difficult to guess what you are talking about.

"a timeSeries object (SC.Adjusted) with 12 variables" doesn't tell me 
whether you have a returns object with 12 columns, where each column 
represents the returns of one asset or whether you have twelve 
observations on a single return series.

If you have a univariate series, as the documentation says, you will use 
the skewness of the series as m3, see ?skewness

If you have a multivariate series, and you're doing component ES, you 
will again follow the documentation and use the coskewness matrix.  see 
?CoSkewnessMatrix

If you want the math, it is also all in the documentation (I suggest 
looking at the pdf version so that the math will be laid out nicely) and 
in even more detail in the references.

Given that R doesn't run on an iPhone, perhaps next time you would be so 
kind as to construct a reproducible example on your computer and use the 
browser interface to gmail to provide a bit more detail on your query 
(providing your name wouldn't hurt either).

Regards,

    - Brian

-- 
Brian G. Peterson
http://braverock.com/brian/
Ph: 773-459-4973
IM: bgpbraverock