Fit ratio model using HMC/NUTS

Usage

fit_hmc(
  formula,
  data,
  family,
  spatial = NULL,
  temporal = NULL,
  spatiotemporal = NULL,
  zi = NULL,
  latent = NULL,
  priors = NULL,
  iter = 2000,
  warmup = floor(iter/2),
  chains = 4,
  cores = NULL,
  L = 0,
  adapt_delta = NULL,
  max_treedepth = NULL,
  seed = NULL,
  verbose = TRUE,
  gradient_mode = "auto",
  re_param = "noncentered",
  metric = "auto",
  walnuts = FALSE
)

Arguments

formula: A ratiod_formula object
data: Data frame
family: Model family
spatial: Optional spatial structure (spatial_car or spatial_bym2)
temporal: Optional temporal structure (temporal_rw1, temporal_ar1, etc.)
zi: Optional zero-inflation specification (ratiod_zi object)
priors: Prior specification
iter: Total number of iterations per chain
warmup: Number of warmup iterations per chain
chains: Number of chains to run (can be parallelized)
cores: Core budget for the fit. Split as concurrent_chains = min(cores, chains) chains in flight at once, each running with cores %/% concurrent_chains threads for its own within-chain gradient/likelihood work. Defaults to the machine's OpenMP thread cap.
L: Number of leapfrog steps per iteration
seed: Random seed for reproducibility
verbose: Print progress
gradient_mode: Gradient computation method: "auto", "H", "A_r", "A", "N"
re_param: Random effects parameterization: "centered" or "noncentered"
metric: Mass matrix type: "auto" (default), "diag", "dense", or "block_diag"
walnuts: Run each NUTS trajectory as a WALNUTS transition: TRUE or FALSE (default)

Value

A ratiod_fit object