Compare runtime using CGGibbs and naive approach to calculate linear predictor — compare_eta_comptime_across

The comparison of the methods is described in this vignette. The user can specify different arguments to alter the specification of the comparison, but it's possible to only specify values of the n_vars argument. It's not possible to change the fact that data is simulated with a normal response and normal explanatory variables.

Usage

compare_eta_comptime_across_nvars(
  n_vars,
  n = 100,
  beta_prior = distributional::dist_normal(0, 1),
  log_likelihood_extra_args = list(sd = 1),
  sample_method = c("slice_sampling", "normal-normal"),
  qslice_fun = qslice::slice_stepping_out,
  ...,
  n_samples = 500,
  burnin = 100,
  parallelise = FALSE,
  n_cores = as.numeric(Sys.getenv("NUMBER_OF_PROCESSORS")) - 1
)

Arguments

n_vars: a numeric vector with numbers of normally distributed columns in the design matrix for different runs
n: a numeric with the number of observations in the data analysed by the GLM
beta_prior: a distribution object created by a function from the distributional package. Could fx. be distributional::dist_normal(mean = 0, sd = 1).
log_likelihood_extra_args: a named list with arguments passed onto the log_density function. Fx. specification of log_likelihood_extra_args = list(sd = x) is needed for the case of family = "gaussian"
sample_method: a character specifying the method used for sampling. The default "slice_sampling" is the intended value in most cases, as it works for any specification of family and beta_prior. "normal-normal" uses a conditional normal distribution to sample from in case of conjugate prior with gaussian response and beta_prior. Implemented for testing purposes but works for that niche case.
qslice_fun: a function from the qslice package. Default is qslice::slice_stepping_out which uses the slice sampler from Neal 2003, but all functions are available.
...: arguments passed onto the function specified by qslice_fun. For default qslice::slice_stepping_out w needs to be specified, while for fx. qslice::slice_elliptical, mu and sigma need to be specified
n_samples: a numeric with number of samples to draw of each parameter(/variable) in the model
burnin: a numeric with the number of samples to be marked as "burnin". Burnin samples are not included in the beta_mean calculation to increase finite sample performance of the LLN estimate
parallelise: a logical if the runs of the algorithm across values of n_vars should be parallelised using future.apply::future_lapply
n_cores: a numeric with the number of cores to use for parallelisation. Default is 1 less than the number of available cores

Value

a data.frame with information on computation time for different values of n_vars

Examples

# Compare the runtime for 2, 20, 40, 60 variables in the model
compare_eta_comptime_across_nvars(n_vars = c(2, seq(from = 20, to = 60, by = 20)),
                                  n_samples = 1,
                                  burnin = 0)
#>              time linear_predictor_calc n_vars n_samples beta_mean
#> 1 0.01716685 secs                update      2         1         0
#> 2 0.02163410 secs                 naive      2         1         0
#> 3 0.46144390 secs                update     20         1         0
#> 4 0.20999956 secs                 naive     20         1         0
#> 5 0.84110141 secs                update     40         1         0
#> 6 1.39404273 secs                 naive     40         1         0
#> 7 1.70097470 secs                update     60         1         0
#> 8 1.65817809 secs                 naive     60         1         0
#>   beta_variance   family sd                 qslice_fun   w parallelised
#> 1             1 gaussian  1 qslice::slice_stepping_out 0.5        FALSE
#> 2             1 gaussian  1 qslice::slice_stepping_out 0.5        FALSE
#> 3             1 gaussian  1 qslice::slice_stepping_out 0.5        FALSE
#> 4             1 gaussian  1 qslice::slice_stepping_out 0.5        FALSE
#> 5             1 gaussian  1 qslice::slice_stepping_out 0.5        FALSE
#> 6             1 gaussian  1 qslice::slice_stepping_out 0.5        FALSE
#> 7             1 gaussian  1 qslice::slice_stepping_out 0.5        FALSE
#> 8             1 gaussian  1 qslice::slice_stepping_out 0.5        FALSE