Re: [Quantlib-users] BlackKarasinski Class in Quantlib

[Peter C. <pca...@gm...>]

Hi Brian,

a and alpha refer to the same variable, namely the reversion speed of
the model, we should probably make the documentation clearer here. The
parameter theta(t) on the other hand is automatically calibrated to
the initial rate curve, i.e. you can not (don't need to) specify this
parameter explicitly.

Best Regards
Peter

Peter

On Sun, 25 Oct 2020 at 09:46, Brian Smith <bri...@gm...> wrote:
>
> Hi,
>
> I am confused about the definition of the Black Karasinski Class
> implemented in Quantlib.
>
> The reference is available in
> https://rkapl123.github.io/QLAnnotatedSource/d4/d58/class_quant_lib_1_1_black_karasinski.html
>
> As per this documentation, the constructor of this class takes a
> parameter called "a" which is defaulted at 0.1
>
> But in this same documentation, the actual process is defined as -
>
> \[ d\ln r_t = (\theta(t) - \alpha \ln r_t)dt + \sigma dW_t, \]
>
> Can you please help me understand what is the relation between a and
> {theta, alpha} in above two forms of specification?
>
> Thanks for your time. Regards
>
>
> _______________________________________________
> QuantLib-users mailing list
> Qua...@li...
> https://lists.sourceforge.net/lists/listinfo/quantlib-users