Parameter estimation of a regime-switching model using an inverse stieltjes moment approach

We address the problem of recovering the time-dependent parameters

of the Black-Scholes option pricing model when the underlying stock

price dynamics are modelled by a finite-state, continuous-time Markov

chain. The coupled system of Dupire-type partial differential equations

is derived and formulated as an inverse Stieltjes moment problem. We

provide numerical illustration on how to apply our method to simulated

financial data. The accuracy of the model parameter estimation is examined

and sensitivity analyses are included to study the behaviour of

the estimated results when model parameters are varied.