Subordinated Wright-Fisher Priors
Nathan A. Judd, Dario Span\`o

TL;DR
This paper introduces a novel class of time-dependent Dirichlet priors based on subordinated Wright-Fisher diffusions, enabling discontinuities and non-Markovian memory, with explicit sampling algorithms for Bayesian inference.
Contribution
It generalizes Wright-Fisher priors through stochastic time-change, providing explicit representations, sampling algorithms, and a new class of discrete dual processes for Bayesian analysis.
Findings
Explicit sampling algorithms for prior and posterior distributions.
A new class of discrete dual processes for computation and conjugacy.
Demonstrates the flexibility of the subordinated Wright-Fisher priors in Bayesian models.
Abstract
A new class of time-dependent Dirichlet priors is introduced as a generalisation of the Wright-Fisher diffusion, allowing discontinuities in the trajectories, as well as non-Markovian memory. This class is obtained as a simple stochastic time-change (subordination), interpreted as a hyper-prior assigned to the operational time-clock of a Wright-Fisher diffusion. Explicit representations and exact sampling algorithms are obtained for prior and posterior distributions of the process and of its clock, given partially exchangeable data sampled at discrete time-points. Computability and conjugacy rely on a novel class of discrete dual processes, generalising existing results on duality and computable filters.
Peer Reviews
No public reviews on file for this paper yet. If you reviewed it on a platform where reviews are public (OpenReview, ICLR, NeurIPS, ICML), you can paste yours below so the community can read it here.
Videos
No videos yet. Explain this paper in a talk, walkthrough, or lecture? Add one.
