R help - Jan 2013 - fPortfolio-portfolio optimization

Hello everyone,

I have been spending many hours on a seemingly simple portfolio optimization
problem using the package fPortfolio.

My optimization problem is slightly different than a standard one such that I
have a known set of asset returns. My problem is how to collect this information
into my functions and pass them onto the optimization function.

I have written my own covariance estimation function using the
"shrinkEstimator" as template. I will use the shrunk estimation of the
covariance matrix with my own set of predicted returns. My code is below.

Many thanks,
Darius

-----------------------------------------------------------------------
b=ts(ret.forecast[1,])

mu.pred=b

myEstimator=function(x) 
{
    stopifnot(inherits(x, "timeSeries"))
    x.mat = x
    mu = mu.pred
    Sigma = .cov.shrink(x = x.mat, verbose = FALSE, ...)
    attr(Sigma, "lambda.var") <- NULL
    attr(Sigma, "lambda.var.estimated") <- NULL
    list(mu = mu, Sigma = Sigma)
} 

portfolio1=portfolioSpec()

a=ts(ret.mat[(1:60),(1:n.assets)])

setEstimator(portfolio1)="myEstimator"

portfolio2=tangencyPortfolio(data=a, spec=portfolio1)

 		 	   		  
	[[alternative HTML version deleted]]