The paper use Markov Chain Monte Carlo (MCMC) method to estimate the parametersof jump diffusion process mode, it solves the problem that maximum likelihood functionis unbounded. And then with the simulated data to check MCMC algorithm’s efficiency,the results shows that the MCMC is quite stable, which is not sensitive when prior settingchanges. Last the paper uses the data from stock and exchange rate to check MCMC’sempiricism, and the results shows that it is very stable, Thus MCMC algorithm in theestimation of jump diffusion random process model is reliable which make a goodfoundation in analysis the sudden jump of financial property. |