Fitting Financial Models to Market Data Using Kriging

Eelse-jan Stutvoet

Site of the project:
Delft University of Technology

start of the project: October 2006

For working address etc. we refer to our alumnipage.

Summary of the master project:
To obtain an optimal parameter set for various financial models, we minimize the distance of a set to the market data in the least square sense. We use a Kriging model, which approximates the surface of model from sample points. These sample points are treated as realizations of a random variable. This way we obtain a mean and a variance for each parameter set. Using a genetic algorithm we search for the point which is expected to improve the model the most and include this point in the approximation. We repeat this process until the optimal parameter set is found.

Contact information: Kees Vuik

Back to the home page or the Master students page of Kees Vuik