setupExperiment

name

seed

number of simulations to at each parameter
  configuration.

list of number of parameter
  configurations in each design, e.g.
  resolution=list(h=5L, d=3L)

resolution

list of parameter bounds used for
  experimental design, e.g.  bounds=list(h=c(0,1),
  d=c(0.001, 2), s=c(200L, 2000L)) where - h : bounds of
  the heterogeneity.  - d : bounds of the
  heteroscedasticity.  - a : bounds of the balancing factor
  of group assignments.  - s : bounds of the study sizes.
  - r : fixed risk in the control.

bounds

sgnf

piv_draws

further arguments to makeExperimentRegistry,
  e.g.  file.dir=tempfile().


  Sets up a computer experiment evaluating the performance
  of different inference methods in the random effects meta
  regression model.


Provides methods for making inference in the random effects meta
regression model such as point estimates and confidence intervals for the
heterogeneity parameter and the regression coefficients vector.  Inference
methods are based on different approaches to statistical inference.
Methods from three different schools are included: methods based on the
method of moments approach, methods based on likelihood, and methods based
on generalised inference.  The package also includes tools to run extensive
simulation studies in parallel on high performance clusters in a modular
way.  This allows extensive testing of custom inferential methods with all
implemented state-of-the-art methods in a standardised way.  Tools for
evaluating the performance of both point and interval estimates are
provided.  Also, a large collection of different pre-defined plotting
functions is implemented in a ready-to-use fashion.

setupExperiment: Running a computer experiment in batch mode

Description

Usage

Arguments

Value