Title A Monte Carlo EM algorithm for discretely observed Diffusions, Jump-diffusions and Lévy-driven Stochastic Differential Equations
Authors Erik Lindström
Alternative Location http://naun.org/multimedia/...
Publication International journal of mathematical models and methods in applied sciences
Year 2012
Volume 6
Issue 5
Pages 643 - 651
Document type Article
Status Published
Quality controlled Yes
Language eng
Publisher The North Atlantic University Union (NAUN) / World Scientific and Engineering Academy and Society (WSEAS)
Abstract English Stochastic differential equations driven by standard<br> Brownian motion(s) or Lévy processes are by far the most popular<br> models in mathematical finance, but are also frequently used in<br> engineering and science. A key feature of the class of models is<br> that the parameters are easy to interpret for anyone working with<br> ordinary differential equations, making connections between statistics<br> and other scientific fields far smoother.<br> We present an algorithm for computing the (historical probability<br> measure) maximum likelihood estimate for parameters in diffusions,<br> jump-diffusions and Lévy processes. This is done by introducing<br> a simple, yet computationally efficient, Monte Carlo Expectation<br> Maximization algorithm. The smoothing distribution is computed<br> using resampling, making the framework very general.<br> The algorithm is evaluated on diffusions (CIR, Heston), jump-diffusion<br> (Bates) and Lévy processes (NIG, NIG-CIR) on simulated<br> data and market data from S &amp; P 500 and VIX, all with satisfactory<br> results.
Keywords Bates model, Heston model, Jump-Diffusion, Lévy process, parameter estimation, Monte Carlo Expectation Maximization, NIG, Stochastic differential equation,
ISBN/ISSN/Other ISSN: 1998-0140

Questions: webmaster
Last update: 2013-04-11

Centre for Mathematical Sciences, Box 118, SE-22100, Lund. Telefon: +46 46-222 00 00 (vx)