[Rquantlib-devel] Fwd: Project Idea and hope for feedback

[Khanh Nguyen]

(Moving to rquantlive-devel list....)

On Tue, Apr 6, 2010 at 3:59 AM, SaichiuNelson Tong
<tongsnelson.ise at gmail.com> wrote:
> thanks for the reply, Khanh.

No problem...

> Actually I don't really know all that much on things related to quant/math
> finance, there is still so much I want to learn in that area, we can
> definitely learn some of these topics together.

Cool... I am looking forward to it.

> I can see why having access to the various "YieldTermStructure" models from
> quantlib  would make things easier to build.  which discount curves/rate
> structures are already covered in rquantlib ?

It's hard to answer the question exactly...Currently, we have
DiscountCurve and FittedBondCurve that can build the term structure.
Since we can't construct and move C++ objects during run-time in R
(right Dirk?), we have to reconstruct the curve from raw data points
whenever we want to use it. For example

curve <- DiscountCurve(.....)
z <- ZeroCouponBond(bondparameters,.... curve,..)

'curve' is a yield term structure constructed from market quotes using
DiscountCurve(). It contains 'table' - a two column data frame with
raw data points for, if I recall correctly, 50 years... When we want
to use the curve to price a zero coupon bond, we reconstruct it from
the data point using QuantLib. Once we get the structure, we can
construct other objects whose constructor asks for a
'YieldTermStructure' object, like the Ibor objects..

There was a small mistype in the last email . Sorry, it was late..:(
and my head wasn't 100% clear... I actually meant the 'term structure'
instead of just 'yield term structure'... The
'VolatilityTermStructure', it is another common parameters for some
important stochastic processes. Having some codes to input and move
them around will be really helpful. Well... that's just my morning
idea...

Cheers,

Khanh Nguyen