Inference for non-linear diffusions and jump-diffusions: a Monte Carlo EM approach
Erik Lindström · Lund University Publications (Lund University) · 2012
We propose a simple, general and computationally efficient algorithm for maximum likelihood estima- tion (MLE) of parameters in diffusion and jump-diffusion processes. This is conducted within a Monte Carlo EM-algorithm, where the smoothing distribution is computed using resampling. The results are encouraging as we can approximate the MLE well for the models studied when using simulated data. We also obtain reasonable estimates, compared to other papers, when fitting the Heston and Bates model to S&P 500 and VIX data.