Home > Research > Publications & Outputs > Simulation based composite likelihood

Links

Text available via DOI:

View graph of relations

Simulation based composite likelihood

Research output: Contribution to Journal/MagazineJournal articlepeer-review

E-pub ahead of print

Standard

Simulation based composite likelihood. / Rimella, Lorenzo; Jewell, Chris; Fearnhead, Paul.
In: Statistics and Computing, Vol. 35, No. 3, 58, 30.06.2025.

Research output: Contribution to Journal/MagazineJournal articlepeer-review

Harvard

Rimella, L, Jewell, C & Fearnhead, P 2025, 'Simulation based composite likelihood', Statistics and Computing, vol. 35, no. 3, 58. https://doi.org/10.1007/s11222-025-10584-z

APA

Rimella, L., Jewell, C., & Fearnhead, P. (2025). Simulation based composite likelihood. Statistics and Computing, 35(3), Article 58. Advance online publication. https://doi.org/10.1007/s11222-025-10584-z

Vancouver

Rimella L, Jewell C, Fearnhead P. Simulation based composite likelihood. Statistics and Computing. 2025 Jun 30;35(3):58. Epub 2025 Feb 25. doi: 10.1007/s11222-025-10584-z

Author

Rimella, Lorenzo ; Jewell, Chris ; Fearnhead, Paul. / Simulation based composite likelihood. In: Statistics and Computing. 2025 ; Vol. 35, No. 3.

Bibtex

@article{8e0661a198f04343b72ed4b4d7dc03a1,
title = "Simulation based composite likelihood",
abstract = "Inference for high-dimensional hidden Markov models is challenging due to the exponential-in-dimension computational cost of calculating the likelihood. To address this issue, we introduce an innovative composite likelihood approach called “Simulation Based Composite Likelihood” (SimBa-CL). With SimBa-CL, we approximate the likelihood by the product of its marginals, which we estimate using Monte Carlo sampling. In a similar vein to approximate Bayesian computation (ABC), SimBa-CL requires multiple simulations from the model, but, in contrast to ABC, it provides a likelihood approximation that guides the optimization of the parameters. Leveraging automatic differentiation libraries, it is simple to calculate gradients and Hessians to not only speed up optimization but also to build approximate confidence sets. We present extensive empirical results which validate our theory and demonstrate its advantage over SMC, and apply SimBa-CL to real-world Aphtovirus data.",
keywords = "Hidden Markov model, Monte Carlo approximation, Individual-based models, Composite likelihood",
author = "Lorenzo Rimella and Chris Jewell and Paul Fearnhead",
year = "2025",
month = feb,
day = "25",
doi = "10.1007/s11222-025-10584-z",
language = "English",
volume = "35",
journal = "Statistics and Computing",
issn = "0960-3174",
publisher = "Springer Netherlands",
number = "3",

}

RIS

TY - JOUR

T1 - Simulation based composite likelihood

AU - Rimella, Lorenzo

AU - Jewell, Chris

AU - Fearnhead, Paul

PY - 2025/2/25

Y1 - 2025/2/25

N2 - Inference for high-dimensional hidden Markov models is challenging due to the exponential-in-dimension computational cost of calculating the likelihood. To address this issue, we introduce an innovative composite likelihood approach called “Simulation Based Composite Likelihood” (SimBa-CL). With SimBa-CL, we approximate the likelihood by the product of its marginals, which we estimate using Monte Carlo sampling. In a similar vein to approximate Bayesian computation (ABC), SimBa-CL requires multiple simulations from the model, but, in contrast to ABC, it provides a likelihood approximation that guides the optimization of the parameters. Leveraging automatic differentiation libraries, it is simple to calculate gradients and Hessians to not only speed up optimization but also to build approximate confidence sets. We present extensive empirical results which validate our theory and demonstrate its advantage over SMC, and apply SimBa-CL to real-world Aphtovirus data.

AB - Inference for high-dimensional hidden Markov models is challenging due to the exponential-in-dimension computational cost of calculating the likelihood. To address this issue, we introduce an innovative composite likelihood approach called “Simulation Based Composite Likelihood” (SimBa-CL). With SimBa-CL, we approximate the likelihood by the product of its marginals, which we estimate using Monte Carlo sampling. In a similar vein to approximate Bayesian computation (ABC), SimBa-CL requires multiple simulations from the model, but, in contrast to ABC, it provides a likelihood approximation that guides the optimization of the parameters. Leveraging automatic differentiation libraries, it is simple to calculate gradients and Hessians to not only speed up optimization but also to build approximate confidence sets. We present extensive empirical results which validate our theory and demonstrate its advantage over SMC, and apply SimBa-CL to real-world Aphtovirus data.

KW - Hidden Markov model

KW - Monte Carlo approximation

KW - Individual-based models

KW - Composite likelihood

U2 - 10.1007/s11222-025-10584-z

DO - 10.1007/s11222-025-10584-z

M3 - Journal article

VL - 35

JO - Statistics and Computing

JF - Statistics and Computing

SN - 0960-3174

IS - 3

M1 - 58

ER -