/marglike

Computation of the marginal likelihood from MCMC output

GNU General Public License v3.0GPL-3.0

Computation of the marginal likelihood from MCMC output (marglike)

This repository will contain code to estimate the marginal likelihood, that is Pr(data|model), from the output of a MCMC sampler.

This technique has been proposed in Robert & Wraith (2009), and further developed in Acerbi et al. (2018).

References

  1. Robert CP, Wraith D. (2009). Computational methods for Bayesian model choice. In P. M. Goggans, & C. Y. Chan (Eds.), AIP Conference Proceedings (Vol. 1193, No. 1, pp. 251-262). AIP.
  2. Acerbi L*, Dokka K*, Angelaki DE, Ma WJ (2018). Bayesian comparison of explicit and implicit causal inference strategies in multisensory heading perception. PLoS Computational Biology 14(7): e1006110. DOI: 10.1371/journal.pcbi.1006110 (*equal contribution)

License

marglike is released under the terms of the GNU General Public License v3.0.