Likelihood-free Bayesian inference for dynamic, stochastic simulators in the social sciences

Thesis

Simulation models – such as agent-based models (abms) in the social sciences – are now used widely across scientific and commercial domains. However, such models often lack a tractable likelihood function, precluding standard likelihood-based statistical inference. In response to this challenge, the past two decades have seen the development of likelihood-free, simulation-based procedures for inferring simulator parameters within the computational statistics and machine learning communitie...

Expand abstract

Authors: Dyer, J

• DOI: 10.5287/ora-14eg5jyqe
• Type of award: DPhil
• Level of award: Doctoral
• Awarding institution: University of Oxford
• Language: English
• Keywords: time series; density estimation; Bayesian inference; stochastic simulators; probabilistic machine learning; deep learning; machine learning; agent-based models; path signatures
• Subjects: Computational social science; Agent-based models; Bayesian inference; Machine learning; Density estimation; Computational statistics; Computer simulation