bartz.mcmcloop

Functions that implement the full BART posterior MCMC loop.

Running the MCMC

run_mcmc(key, state, n_save, *[, n_burn, ...])	Run the MCMC for the BART posterior.
RunMCMCResult(final_state, burnin_trace, ...)	Return value of run_mcmc.
BurninTrace(has_chains, mesh, ...)	MCMC trace with only diagnostic values.
MainTrace(has_chains, mesh, grow_prop_count, ...)	MCMC trace with trees and diagnostic values.
MainTraceWithTrainPred(has_chains, mesh, ...)	Main trace that also stores the latent predictions at the training points.

Evaluating the trace

evaluate_trace(X, trace, *[, ...])	Compute predictions for all iterations of the BART MCMC.
compute_varcount(p, trace, *[, out_chain_axis])	Count how many times each predictor is used in each MCMC state.
EvaluableTrace(*args, **kwargs)	Structural type of the traces accepted by evaluate_trace.

Progress callbacks

The entry points are make_print_callback and make_tqdm_callback.

make_print_callback(state, *[, dot_every, ...])	Prepare a progress-printing callback for run_mcmc.
make_tqdm_callback(state, *[, update_every, ...])	Prepare a tqdm progress-bar callback for run_mcmc.
Callback()	Base class for run_mcmc callbacks.
CallbackTuple(callbacks)	Callback that runs a sequence of callbacks in order.
CheckPlatformCallback(platform)	Check the given platform matches the actual platform at runtime.
PrintCallback(dot_every, report_every, ...)	Progress-printing callback for run_mcmc, see make_print_callback.
TqdmCallback(bar_id, update_every, ...)	tqdm progress-bar callback for run_mcmc, see make_tqdm_callback.
StatsAccumulator(sums, count)	Running average of the forest diagnostics shown during the MCMC.
StatsReport(grow_prop, move_acc, ...)	Forest diagnostics produced by StatsAccumulator.report for one report.