Does anyone have experience about Heston model's calibration?

  • Thread starter Thread starter raywin
  • Start date Start date
R

raywin

Joined
1/4/08
Messages
24
Points
11
Hello Every

Does anyone have experience about Heston model's calibration? I just followed nimalin moodley 's paper. But I can not get the same results as page 29. Did anyone do that before? Thanks a lot!


Ray
 
Hello Ray,

a friend of mine has compute a Heston-Calibration-program in C++.
But I do not know his experience with the nimalin moodley 's paper.

I ask him for further information.


But in the meantime here is a paper about Heston Calibration and a Matlab-Code:
 

Attachments

Yes, I try that code. But I think that is not really correct. The main idea is correct but there are some problems. But I can not get the result in page 29 of that paper.
 
Thanks. I saw this code before in some web. Currently I just want to use mablab build-in function lsqnonlin to do calibration.
 
The idea is not really hard but there are some trivial work to do. I am researching that. Hope crack it asap!
 
Dear all,
I wonder whether it is logical to look at the term structure of the parameters in Heston model.
I mean do you think it is logical to calibrate Heston model to a spesific maturity on a spesific day(and doing this seperately for every maturity) and looking at evolution of the term structure of parameters by following the same procedure.
Do you have any experience about this?
Your help will be very appreciated.
 
mixumus said:
Dear all,
I wonder whether it is logical to look at the term structure of the parameters in Heston model.
I mean do you think it is logical to calibrate Heston model to a spesific maturity on a spesific day(and doing this seperately for every maturity) and looking at evolution of the term structure of parameters by following the same procedure.
Do you have any experience about this?
Your help will be very appreciated.
Click to expand...

Not sure if I understand well your question. You could use Heston with piecewise constant parameterization to calibrate a given smile surface.
 
raywin said:
The idea is not really hard but there are some trivial work to do. I am researching that. Hope crack it asap!
Click to expand...

what's your problem exactly? I haven't seen any result on page 29, except a graph of error calibration with no numerical output.

Calibration of Heston is quite tricky though. It's not as easy as in Moodley's. You should use Gauss-Lombarto and try to use Lipton-Lewis to save time of computation as you'll have to compute only 1 integral instead of 2.

Hope it helps.
 
You should use a global optimization technique (e.g. SA) to calibrate the model since you could end up in a local minima when using nonlinear least squares method. The optimization problem is not convex.
 
The Levenberg-Marquardt Method should also be recommendable.
 
post the question in nuclear phynance
 
MagicWeasel said:
You should use a global optimization technique (e.g. SA) to calibrate the model since you could end up in a local minima when using nonlinear least squares method. The optimization problem is not convex.
Click to expand...

using fminsearch in Matlab will do the trick

I got good calibration results not only for Heston, but also for the more advanced Heston-Nandi (2000) model
 
hey every one
.I am doing my master thesis in volatility forecating but I have a problem. I want to calibrate heston model as discribed in the following articles but I really cant find any thing (any code or help). I am trying hard but I am not good at programming so its tough. Maybe u or any body else on this site help me with the co de ? thanks in advance:)



http://www.wilmott.com/pdfs/040502_shu.pdf

http://www.google.com/url?sa=t&rct=j&q=pricing s&p 500 index options with heston’s model&source=web&cd=2&ved=0CC4QFjAB&url=http://citeseerx.ist.psu.edu/viewdoc/download?doi=10.1.1.201.590&rep=rep1&type=pdf&ei=KoOiT8WkB82k4AStnJWuCQ&usg=AFQjCNE2oGOjaOBlfPX9VQJLTt8tIqVFew&cad=rja
 

Similar threads

BBB
Anyone have experience using any website designers?
Replies
3
Views
2K
MRoss
MRoss
Back
Top Bottom