R-package agentBayes: Likelihood-based statistical methods for agent-based models
6 affiliations
- 1Department of Biological and Environmental Science, University of Jyväskylä, Jyväskylä, Finland
- 2Department of Mathematics, Swansea University, Swansea, United Kingdom
- 3Department of Evolution, Ecology, and Behaviour, Institute of Infection, Veterinary, and Ecological Sciences, University of Liverpool, Liverpool, United Kingdom
- 4Department of Information Technology, Uppsala University, Uppsala, Sweden
- 5Department of Translational Hematology and Oncology Research, Cleveland Clinic, Cleveland, Ohio, United States of America
- 6Department of Medical Genetics, Oslo University Hospital, Oslo, Norway
Abstract
Statistically analysing interacting particle systems remains challenging because the governing equations are analytically intractable. Existing solutions include moment closure methods with pseudolikelihood-based frameworks, and likelihood-free frameworks based on extensive simulations, both relying on heuristic choices whose validity is difficult to predict. As a resolution, we rigorously derive an asymptotically exact expression for the likelihood of agent-based models (ABMs) operating in continuous space and time that can be formulated as reactant–catalyst–product (RCP) models. We derive an expression for the conditional density of agents given information about the current and earlier distributions of neighbouring agents. We utilize this expression to construct an asymptotically exact likelihood that applies to both spatial snapshot and time-series data. We implement the likelihood expression and a Bayesian parameter estimation framework in the R-package agentBayes and demonstrate its utility in biological research and beyond with simulated case studies and empirical data on the evolution of cancer cell populations. Author summary: Many study systems across biology, finance, physics and the social sciences consist of discrete interacting agents, yet fitting agent-based models to data remains challenging as most models are too complicated to be treated analytically. We resolve this fundamental challenge and derive tractable likelihoods for a broad class of agent-based models and data, such as spatial snapshot and time-series data. Accompanied by our statistical software, we show how to apply our framework to address biological research questions at the example of the coevolution of cancer cell populations.
Related
- Contrast-Aware Annotation for Statistical Graphics: The ggtwotone Package for ggplot2
- A Decision-Theoretic Framework for Comparing Likelihood Ratio Methods for the Rare Type Match Problem
- Strategic Multi-Agent Resource Allocation via Statistical Contracts
- Riemannian Newton Methods with Statistical Inference
- Finite difference methods for a continuous-time heterogeneous agent model with recursive utility
BibTeX
@article{moser2026rpackage,
title = {{R-package agentBayes: Likelihood-based statistical methods for agent-based models}},
author = {Niklas Moser and Dmitri Finkelshtein and Georgy Chargaziya and Stephen J. Cornell and Sara Hamis and Jacob G. Scott and Dagim Shiferaw Tadele and Otso Ovaskainen},
journal = {PLOS Computational Biology},
year = {2026},
month = sep,
doi = {10.1371/journal.pcbi.1014786},
url = {https://journals.plos.org/ploscompbiol/article?id=10.1371/journal.pcbi.1014786},
note = {CC BY 4.0},
}
RIS
TY - JOUR
TI - R-package agentBayes: Likelihood-based statistical methods for agent-based models
AU - Niklas Moser
AU - Dmitri Finkelshtein
AU - Georgy Chargaziya
AU - Stephen J. Cornell
AU - Sara Hamis
AU - Jacob G. Scott
AU - Dagim Shiferaw Tadele
AU - Otso Ovaskainen
PY - 2026
DA - 2026/09/15
JO - PLOS Computational Biology
DO - 10.1371/journal.pcbi.1014786
UR - https://journals.plos.org/ploscompbiol/article?id=10.1371/journal.pcbi.1014786
AB - Statistically analysing interacting particle systems remains challenging because the governing equations are analytically intractable. Existing solutions include moment closure methods with pseudolikelihood-based frameworks, and likelihood-free frameworks based on extensive simulations, both relying on heuristic choices whose validity is difficult to predict. As a resolution, we rigorously derive an asymptotically exact expression for the likelihood of agent-based models (ABMs) operating in continuous space and time that can be formulated as reactant–catalyst–product (RCP) models. We derive an expression for the conditional density of agents given information about the current and earlier distributions of neighbouring agents. We utilize this expression to construct an asymptotically exact likelihood that applies to both spatial snapshot and time-series data. We implement the likelihood expression and a Bayesian parameter estimation framework in the R-package agentBayes and demonstrate its utility in biological research and beyond with simulated case studies and empirical data on the evolution of cancer cell populations. Author summary: Many study systems across biology, finance, physics and the social sciences consist of discrete interacting agents, yet fitting agent-based models to data remains challenging as most models are too complicated to be treated analytically. We resolve this fundamental challenge and derive tractable likelihoods for a broad class of agent-based models and data, such as spatial snapshot and time-series data. Accompanied by our statistical software, we show how to apply our framework to address biological research questions at the example of the coevolution of cancer cell populations.
SN - 1553-7358
ER -