Package: fBonds
Yohan Chalabi
chalabi at phys.ethz.ch
Wed Sep 23 00:09:38 CEST 2009
>>>> "SG" == "Sergio Guirreri" <sergioguirreri at gmail.com>
>>>> on Sun, 26 Jul 2009 17:16:06 +0200
SG> I'm using the package fBonds to estimate the yield curve
SG> by Nelson
SG> Siegel's method. I'm focusing on the code of the function
SG> NelsonSiegel
SG> and I note that could be a problem:
SG>
SG> 1)
SG> the graphic of the yield curve is done with the vector of
SG> the starting
SG> parameter and not whit the estimated parameters. I report the
SG> piece of
SG> code:
SG> fit = nlminb(start = gvec, objective = func, rate = rate,
SG> maturity = maturity, control = list(eval.max = 2000,
SG> iter.max = 2000))
SG> names(fit) = c(beta0, beta1, beta2, tau1)
SG> if (doplot) {
SG> ---->>> yfit = fx(maturity, gvec)
SG> ---->>> plot(maturity, rate, ylim = c(min(c(rate, yfit)),
SG> max(c(rate,
SG> yfit))), main = Nelson-Siegel)
SG> ---->>> lines(maturity, yfit, col = steelblue)
SG> grid()
SG>
SG> the object yfit contains the interest rates estimated with
SG> the vector
SG> gvec which is the starting parameter, while the estimated
SG> coefficients
SG> are in the object fit.
SG>
SG> Could the same problem be in the function Svensson ?
SG>
SG> Best regards Sergio Guirreri
SG>
Hi Sergio,
Thanks for the report. I changed the code in fBonds.
Regards,
Yohan
--
PhD student
Swiss Federal Institute of Technology
Zurich
www.ethz.ch
More information about the Rmetrics-core
mailing list