[R-SIG-Finance] FPortfolio / MAxReturnPortfolio

[u0055 at wolke7.net]

Alexios,

Thanks a lot for your very fast reaction.
I installed the new parma_1.5-2 successfully and
it's running without any problems :-)

Uwe


> ok, thats a bug which defaults to targetType='inequality' irrespective
> of value passed.
>
> 1. Download latest (fixed) version from either r-forge (you'll have to
> wait until it rebuilds) or immediately from bitbucket (assuming you can
> build from source):
> library(devtools)
> install_bitbucket("parma","alexiosg")
>
> 2. Make sure to set targetType='equality'.
>
> Regards,
>
> Alexios
>
> PS Try to create a minimally reproducible example next time which does
> not need so much data to be included in the email.
>
> On 09/07/2014 11:16, u0055 at wolke7.net wrote:
>> Thanks Alexios,
>>
>> parma is great and
>>
>> #############
>> library(parma)
>> c = cov( datamatrix )
>> if( ! is.positive.definite( c )) c = make.positive.definite( c )
>> spec = parmaspec( S=c, risk="EV", riskType="minrisk", target=0.000057,
>> forecast=colMeans( datamatrix ))
>> weights(parmasolve( spec, solver="SOCP" ))
>> #############
>>
>> worked out for me.
>> The only issue with parma is,
>> that the target set in parmaspec()
>> is not always equal to the "Reward" in the resulting "PARMA Portfolio".
>> Please find my attached example,
>> where the target was set to 0.000057 and
>> parmareward is 0.0008336 .
>>
>> Shouldn't the target return of the optimized portfolio
>> be equal to the parmareward (parmareward:
>> "Extracts the expected reward of the optimized portfolio"),
>> or did I get something wrong ?
>>
>> Thanks in advance,
>> Uwe