addHMC {nimbleHMC}

Add HMC sampler

Description

Add a No-U-Turn (NUTS) Hamiltonian Monte Carlo (HMC) sampler to an existing nimble MCMC configuration object

Usage

addHMC(
  conf,
  target = character(),
  type = "NUTS",
  control = list(),
  replace = FALSE,
  print = TRUE,
  ...
)

Arguments

conf

A nimble MCMC configuration object, as returned by 'configureMCMC'.

target

A character vector of continuous-valued stochastic node names to sample. If this argument contains any discrete-valued nodes, an error is produced and no sampler is added.

type

A character string specifying the type of HMC sampler to add, either "NUTS" or "NUTS_classic". See 'help(NUTS)' or 'help(NUTS_classic)' for details of each sampler. The default sampler type is "NUTS".

control

Optional named list of control parameters to be passed as the 'control' argument to the HMC sampler. See 'help(NUTS)' or 'help(NUTS_classic)' for details of the control list elements accepted by each sampler.

replace

Logical argument. If 'TRUE', any existing samplers operating on the specified nodes will be removed, prior to adding the HMC sampler. Default value is 'FALSE'.

print

Logical argument whether to print the newly added HMC sampler. Default value is 'TRUE'.

...

Additional named arguments passed through ... will be used as additional control list elements.

Details

This function adds an HMC sampler to an MCMC configuration object. Use this function if you have already created an MCMC configuration and want to add an HMC sampler. Optionally, using 'replace = TRUE', this function will also remove any existing samplers operating on the target node(s).

Either the 'NUTS_classic' or the 'NUTS' sampler can be added. Both implement variants of No-U-Turn HMC sampling, however the 'NUTS' sampler uses more modern adapatation techniques. See 'help(NUTS)' or 'help(NUTS_classic)' for details.

Use 'conf$addSampler' instead if you need more fine-grained control. See 'help(configureMCMC)' in nimble.

Value

Invisibly returns an object of class 'MCMCconf', but this function is primary called for its side effect.

Author(s)

Daniel Turek

See Also

configureHMC buildHMC configureMCMC addSampler sampler_NUTS sampler_NUTS_classic

Examples

code <- nimbleCode({
    b0 ~ dnorm(0, 0.001)
    b1 ~ dnorm(0, 0.001)
    sigma ~ dunif(0, 10000)
    for(i in 1:N) {
        mu[i] <- b0 + b1 * x[i]
        y[i] ~ dnorm(mu[i], sd = sigma)
    }
})

N <- 10
constants <- list(N = N, x = 1:N)
data <- list(y = 1:N)
inits <- list(b0 = 1, b1 = 0.1, sigma = 1)

Rmodel <- nimbleModel(code, constants, data, inits, buildDerivs = TRUE)

## create MCMC configuration object
conf <- configureMCMC(Rmodel, nodes = NULL)

## add HMC sampler operating on all stochastic model nodes
addHMC(conf)

Rmcmc <- buildMCMC(conf)

# Cmodel <- compileNimble(Rmodel)
# Cmcmc <- compileNimble(Rmcmc, project = Rmodel)
# samples <- runMCMC(Cmcmc)