My idea is to work with dynamic Nelson Siegel models(DNS) on my master's thesis. As I am finishing undergraduation this year I started researching on the subject. I wonder what is being discussed in this part of Dynamic Nelson Siegel modeling?
I)The estimation method I´d like to use is the Bayesian. What I could do with the Bayesian method to put me in the state of art of DNS models .
II) I also saw that there is a discussion of terms of arbitrage and non arbitrage conditions. What mathematics needed to understand how to include no-arbitrage condition (Stochastic Calculus/Process)? I confess I did not quite understand the paper (below) reporting this idea.
http://www.frbsf.org/economic-research/files/wp07-20bk.pdf (this one , right?)
That would be the type of Dynamic Affine Nelson Siegel models with non arbitrage conditions?
III) So the families of Nelson Siegel models are divided into:
Nelson Siegel Dynamic
Nelson Siegel Dynamic with non arbitrage conditions
Affine Dynamic Nelson Siegel models with conditions not arbitration
As you can see I'm a little lost.
Anyway, any information could be useful to me.
Thanks a lot.