rjpdmp-package | R Documentation |
Implements various reversible jump piecewise-deterministic Markov Process methods including the ZigZag and Bouncy Particle Sampler with Normal or Spherical velocity distributions (Chevallier, Fearnhead, Sutton 2020, https://arxiv.org/abs/2010.11771).
The package can be used to Generates PDMP trajectories for reversible jump
zigzag_logit
: ZigZag on logistic likelihood problem
zigzag_logit_ss
: ZigZag with subsampling on Logistic likelihood problem
bps_s_logit
: BPS with velocities distributed uniformly on the sphere for a Logistic likelihood problem
bps_n_logit
: BPS with velocities distributed Normally for a Logistic likelihood problem
zigzag_rr
: ZigZag on a robust regression likelihood problem
bps_s_rr
: BPS with velocities distributed uniformly on the sphere for a robust regression likelihood problem
bps_n_rr
: BPS with velocities distributed Normally for a robust regression likelihood problem
Additional functions for plotting, generating samples, calculating posterior means or probabilities of inclusion
plot_pdmp
: Plot marginal densities and joint pairs plots for trajectories and samples of PDMP samplers and optionally MCMC samples for comparison.
plot_pdmp_multiple
: Plots to compare PDMP samplers and optionally MCMC samples.
gen_sample
: Get samples from PDMP trajectories taking a fixed time discretisation.
model_probabilities
: Calculate either marginal probabilities of inclusions or posterior probabilities of specific models.
models_visited
: Count the number of times a model is visited
marginal_mean
: Calculate the marginal mean using PDMP trajectories
cond_mean
: Calculate the mean conditioned on being in a specific model
Extensions to the package are planned.
Add the following code to your website.
For more information on customizing the embed code, read Embedding Snippets.