Approximate maximum likelihood estimation using datacloning ABC
(2016) In Computational Statistics & Data Analysis 105. p.166183 Abstract
 A maximum likelihood methodology for a general class of models is presented, using an approximate Bayesian computation (ABC) approach. The typical target of ABC methods are models with intractable likelihoods, and we combine an ABCMCMC sampler with socalled "data cloning" for maximum likelihood estimation. Accuracy of ABC methods relies on the use of a small threshold value for comparing simulations from the model and observed data. The proposed methodology shows how to use large threshold values, while the number of dataclones is increased to ease convergence towards an approximate maximum likelihood estimate. We show how to exploit the methodology to reduce the number of iterations of a standard ABCMCMC algorithm and therefore reduce... (More)
 A maximum likelihood methodology for a general class of models is presented, using an approximate Bayesian computation (ABC) approach. The typical target of ABC methods are models with intractable likelihoods, and we combine an ABCMCMC sampler with socalled "data cloning" for maximum likelihood estimation. Accuracy of ABC methods relies on the use of a small threshold value for comparing simulations from the model and observed data. The proposed methodology shows how to use large threshold values, while the number of dataclones is increased to ease convergence towards an approximate maximum likelihood estimate. We show how to exploit the methodology to reduce the number of iterations of a standard ABCMCMC algorithm and therefore reduce the computational effort, while obtaining reasonable point estimates. Simulation studies show the good performance of our approach on models with intractable likelihoods such as gandk distributions, stochastic differential equations and statespace models. (Less)
Please use this url to cite or link to this publication:
http://lup.lub.lu.se/record/8410635
 author
 Picchini, Umberto ^{LU} and Anderson, Rachele ^{LU}
 organization
 publishing date
 20160819
 type
 Contribution to journal
 publication status
 published
 subject
 keywords
 Approximate Bayesian computation, Intractable likelihood, MCMC, Statespace model, Stochastic differential equation
 in
 Computational Statistics & Data Analysis
 volume
 105
 pages
 18 pages
 publisher
 Elsevier
 external identifiers

 Scopus:84986575881
 ISSN
 01679473
 DOI
 10.1016/j.csda.2016.08.006
 language
 English
 LU publication?
 yes
 id
 d367ad5ca1ff4d12ad3cf3796b135bdc (old id 8410635)
 alternative location
 http://arxiv.org/abs/1505.06318
 date added to LUP
 20160229 16:14:36
 date last changed
 20161003 07:24:53
@misc{d367ad5ca1ff4d12ad3cf3796b135bdc, abstract = {A maximum likelihood methodology for a general class of models is presented, using an approximate Bayesian computation (ABC) approach. The typical target of ABC methods are models with intractable likelihoods, and we combine an ABCMCMC sampler with socalled "data cloning" for maximum likelihood estimation. Accuracy of ABC methods relies on the use of a small threshold value for comparing simulations from the model and observed data. The proposed methodology shows how to use large threshold values, while the number of dataclones is increased to ease convergence towards an approximate maximum likelihood estimate. We show how to exploit the methodology to reduce the number of iterations of a standard ABCMCMC algorithm and therefore reduce the computational effort, while obtaining reasonable point estimates. Simulation studies show the good performance of our approach on models with intractable likelihoods such as gandk distributions, stochastic differential equations and statespace models. }, author = {Picchini, Umberto and Anderson, Rachele}, issn = {01679473}, keyword = {Approximate Bayesian computation,Intractable likelihood,MCMC,Statespace model,Stochastic differential equation}, language = {eng}, month = {08}, pages = {166183}, publisher = {ARRAY(0x92ff2e8)}, series = {Computational Statistics & Data Analysis}, title = {Approximate maximum likelihood estimation using datacloning ABC}, url = {http://dx.doi.org/10.1016/j.csda.2016.08.006}, volume = {105}, year = {2016}, }