[R-SIG-Finance] PerformanceAnalytics - Sharpes Style Analysis- but with intercept?

[Philipp]

Thanks for the commented Function. So i can understand a little bit more
about it. 

In fact I want to run following constrained Regression. R.f=w0 +
w1*Rs_1+w2*Rs_2+w3*Rs_3 
with the constraints: (1) w1+w2+w3 =1 (2) w1,w2,w3>0. So the Intercept(w0)
is not included in the Constraints. In my case, R.f is a country return,
which I regress on an intercept and Industryreturnbenchmarks. The purpose is
to find a quantity of how well the Industries can explain a country-index.
(Which will be done by Rsquared, The question why an intercept is integrated
(That has been done in some papers, considering this topic), I can not
explain.Maybe because of a better fit). 

If I would consider an intercept I have to fix the Calculation of D and d in
the code of style.QPfit. The purpose is to minimize Var(R.fund-sum[w_i*Rs_i
+w0]). Then I get Var(R.f)(=0)+Var(wiRs_i+w0)-2Cov(Rfund, Rstyle+w0). So in
the Calculation of D and d I get D (=Dmat)=Cov(Rstyle+w0); d=Cov(R.fund,
Rstyle+w0). So my Problem is now to find the right function to minimize
under estimating w, without changing the constraints. 

Thanks,

Best regards

Philipp

--
View this message in context: http://r.789695.n4.nabble.com/PerformanceAnalytics-Sharpes-Style-Analysis-but-with-intercept-tp4125767p4129124.html
Sent from the Rmetrics mailing list archive at Nabble.com.