%0 Journal Article
%T Resampling: an improvement of Importance Sampling in varying population size models
%+ Institut Montpelliérain Alexander Grothendieck (IMAG)
%+ Centre de Biologie pour la Gestion des Populations (UMR CBGP)
%+ Institut de Biologie Computationnelle (IBC)
%+ Institut des Sciences de l'Evolution de Montpellier (UMR ISEM)
%+ Institut de Mathématiques de Marseille (I2M)
%A Merle, Coralie
%A Leblois, Raphaël
%A Rousset, François
%A Pudlo, Pierre
%Z gence Nationale de la Recherche (project IM-Model CORAL.FISH 2010-BLAN-1726-01) and by the Institut National de Recherche en Agronomie (Project INRAStartingGroup‘‘IGGiPop’’
%< avec comité de lecture
%@ 0040-5809
%J Theoretical Population Biology
%I Elsevier
%V 114
%P 70 - 87
%8 2017
%D 2017
%Z 1603.07237
%R 10.1016/j.tpb.2016.09.002
%M 27712980
%K Importance Sampling
%K coalescent.
%K resampling
%K demographic inference
%K population genetics
%K jump Markov process
%Z Mathematics [math]/Statistics [math.ST]
%Z Mathematics [math]/Probability [math.PR]
%Z Statistics [stat]/Methodology [stat.ME]
%Z Statistics [stat]/Applications [stat.AP]
%Z Statistics [stat]/Computation [stat.CO]
%Z Life Sciences [q-bio]/Genetics
%Z Life Sciences [q-bio]/Genetics/Populations and Evolution [q-bio.PE]Journal articles
%X Sequential importance sampling algorithms have been defined to estimate likelihoods in models of ancestral population processes. However, these algorithms are based on features of the models with constant population size, and become inefficient when the population size varies in time, making likelihood-based inferences difficult in many demographic situations. In this work, we modify a previous sequential importance sampling algorithm to improve the efficiency of the likelihood estimation. Our procedure is still based on features of the model with constant size, but uses a resampling technique with a new resampling probability distribution depending on the pairwise composite likelihood. We tested our algorithm, called sequential importance sampling with resampling (SISR) on simulated data sets under different demographic cases. In most cases, we divided the computational cost by two for the same accuracy of inference, in some cases even by one hundred. This study provides the first assessment of the impact of such resampling techniques on parameter inference using sequential importance sampling, and extends the range of situations where likelihood inferences can be easily performed.
%G English
%2 https://hal.science/hal-01270437/document
%2 https://hal.science/hal-01270437/file/resampling-improvement-importance-HAL.pdf
%L hal-01270437
%U https://hal.science/hal-01270437
%~ IRD
%~ CIRAD
%~ EPHE
%~ CNRS
%~ INRIA
%~ UNIV-AMU
%~ INRA
%~ I3M_UMR5149
%~ EC-MARSEILLE
%~ ISEM
%~ INSMI
%~ I2M
%~ I2M-2014-
%~ IMAG-MONTPELLIER
%~ AGROPOLIS
%~ PSL
%~ AGREENIUM
%~ MIPS
%~ INRIA2017
%~ B3ESTE
%~ UNIV-MONTPELLIER
%~ INRIA-AUT
%~ TEST-DEV
%~ INSTITUT-AGRO-MONTPELLIER
%~ INRAE
%~ INRAEOCCITANIEMONTPELLIER
%~ EPHE-PSL
%~ CBGP
%~ UM-2015-2021
%~ TEST3-HALCNRS
%~ TEST4-HALCNRS
%~ INSTITUT-AGRO
%~ TEST5-HALCNRS