Particle methods for maximum likelihood estimation in latent variable models
- First Online:
- Cite this article as:
- Johansen, A.M., Doucet, A. & Davy, M. Stat Comput (2008) 18: 47. doi:10.1007/s11222-007-9037-8
- 269 Downloads
Standard methods for maximum likelihood parameter estimation in latent variable models rely on the Expectation-Maximization algorithm and its Monte Carlo variants. Our approach is different and motivated by similar considerations to simulated annealing; that is we build a sequence of artificial distributions whose support concentrates itself on the set of maximum likelihood estimates. We sample from these distributions using a sequential Monte Carlo approach. We demonstrate state-of-the-art performance for several applications of the proposed approach.