This function defines the default values for adaptive proposal control parameters and MCMC control parameters, while allowing users to specify custom values.
Usage
create_mcmc_params(
stage = "stage1",
initial_vcv_weight = 1,
initial_scaling = 2,
initial_scaling_weight = NULL,
min_scaling = 0.1,
scaling_increment = NULL,
log_scaling_update = TRUE,
acceptance_target = 0.234,
forget_rate = 0.6,
forget_end = Inf,
adapt_end = Inf,
pre_diminish = 40000,
n_steps = 10000,
n_burnin = 0,
n_chains = 4,
n_workers = 4,
n_threads_total = 8
)Arguments
- stage
possible values are "NULL", stage1", "stage2", or "noadapt"
- initial_vcv_weight
Weight for the initial variance-covariance matrix (default = 1).
- initial_scaling
Scaling factor for the proposal (default = 2).
- initial_scaling_weight
Optional weight for initial scaling (default = NULL).
- min_scaling
Minimum scaling factor for the proposal (default = 0.1).
- scaling_increment
Increment for scaling factor (default = NULL).
- log_scaling_update
Logical, should the scaling be updated on a log scale? (default = TRUE).
- acceptance_target
The target acceptance rate for the proposal (default = 0.234).
- forget_rate
Rate at which the proposal forgets past history (default = 0.6).
- forget_end
Time step at which the forgetting stops (default = Inf).
- adapt_end
Time step at which the adaptation stops (default = Inf).
- pre_diminish
Time steps before the diminishing starts (default = 40000).
- n_steps
Total number of MCMC steps (default = 10000).
- n_burnin
Number of burn-in steps (default = 0).
- n_chains
Number of MCMC chains (default = 4).
- n_workers
Number of workers for parallel execution (default = 4).
- n_threads_total
Total number of threads for parallel execution (default = 8).