A model selection framework for better option pricing
This repository will (hopefully) contain documentation and codes about my PhD thesis work.
Short description about the work
The idea is simple. There are many models to put a price on an option contract. The problem with the pricing models is they fail to sustain high performance in a rapidly changing environment such as financial markets, even with frequent parameter updates.
We propose a selection method to pick price estimates from different models so properly that the model selection's pickings result in higher performance than any individual model in its model set.
Once a proper model set is determined, model selection can work with minimal tweaking across different time periods, assets and performance metrics.
TODO
- Including proper documentation and a less academic version of the model selection method.
- Adding understandable comments on model selection codes.
- An
R
package.
##Acknowledgments
This study is supported by Bogazici University Scientific Research Projects (Project Number: 8101).