Options for the SAEM stochastic-approximation covariance in setCov()

Used by setCov(fit, "sa"), which runs the SAEM covariance phase at the fit's estimates, after warm-up iterations unless it continues a SAEM fit's own chains. Every population parameter is held at the fit's estimates throughout (mixture proportions excepted), so the covariance is the one at those estimates.

Usage

saControl(
  nBurn = 100L,
  nEm = 100L,
  nSaCov = 500L,
  seed = 99L,
  warmStart = TRUE
)

Arguments

nBurn, nEm

warm-up iterations that equilibrate the MCMC chains before the covariance phase, for a fit without SAEM chain history (or with warmStart = FALSE)

nSaCov

iterations in the covariance phase; more gives a less noisy covariance

seed

random seed

warmStart

for a SAEM fit, continue the fit's own MCMC chains: the covariance phase starts from the fit's last iteration with no new warm-up iterations. This changes the Monte Carlo path, so the covariance differs from a cold start by Monte Carlo noise. Other fits, and SAEM fits that kept no chain history, always run nBurn/nEm.

Value

saControl object

See also

setCov(), saemControl()

Author

Matt Fidler

Examples

saControl(nSaCov = 1000)
#> $nBurn
#> [1] 100
#> 
#> $nEm
#> [1] 100
#> 
#> $nSaCov
#> [1] 1000
#> 
#> $seed
#> [1] 99
#> 
#> $warmStart
#> [1] TRUE
#> 
#> attr(,"class")
#> [1] "saControl"