Calibration Of The Heston Model

Traditionally, the Heston model has been calibrated using a combination of least squares, options inference and gradient
methods. However, a new calibration technique has recently been developed based on the explicit solution and stochastic
calculus techniques. This new method could greatly simplify and improve the accuracy of the process. The explicit price solution
and the filter used in the calibration problem mentioned above are found to be key for an explicit solution to the Markowitz
problem for one Heston stock and one bond.
In this project, we will focus on self-calibration, where we will only use the stock path itself to calibrate the model, i.e. no option
data will be used. This involves calculations of some parameters and a filtering problem for the remainder. The goals are: 1)
Evaluate various filters for the calibration process. 2) Compare our calibration method to others in the literature. 3) Investigate
improvements to the calibration by including option data observations.

Faculty Supervisor:

Michael Kouritzin

Student:

Partner:

Indian Institute of Science Education and Research Pune

Discipline:

Mathematics

Sector:

Finance and Insurance; Other

University:

University of Alberta

Program: