Approximate inference in state space models with intractable likelihoods using Gaussian process optimisation
2014 (English)Report (Other academic)
We propose a novel method for MAP parameter inference in nonlinear state space models with intractable likelihoods. The method is based on a combination of Gaussian process optimisation (GPO), sequential Monte Carlo (SMC) and approximate Bayesian computations (ABC). SMC and ABC are used to approximate the intractable likelihood by using the similarity between simulated realisations from the model and the data obtained from the system. The GPO algorithm is used for the MAP parameter estimation given noisy estimates of the log-likelihood. The proposed parameter inference method is evaluated in three problems using both synthetic and real-world data. The results are promising, indicating that the proposed algorithm converges fast and with reasonable accuracy compared with existing methods.
Place, publisher, year, edition, pages
2014. , 25 p.
LiTH-ISY-R, ISSN 1400-3902 ; 3075
Approximate Bayesian computations, Gaussian process optimisation, Bayesian parameter inference, alpha-stable distribution
Probability Theory and Statistics Control Engineering Signal Processing
IdentifiersURN: urn:nbn:se:liu:diva-106198ISRN: LiTH-ISY-R-3075OAI: oai:DiVA.org:liu-106198DiVA: diva2:714542
FunderSwedish Research Council, 621-2013-5524