get_analysis_mode()
function to get current package optionsImplement "pmb_code"
class to support a strict subset of the nimble language (#64).
check_model_pars()
returns derived
(#62).
engine = "jags"
.
Use raw GitHub file.
Get JAGS.exe from S3 bucket.
Relax test.
param_type
and scalar
arguments from pars.character()
and pars.mb_code()
and scalar
from pars.model()
and add "raw"
parameter type to pars.mb_model()
and pars.mb_analysis().new_expr_vec
to be getOption("mb.new_expr_vec", FALSE)
new_expr_vec = FALSE
to predict()
.new_expr_vec
to be flag that FALSE by default.new_expr = FALSE
argument to mcmc_derive.mb_analyses.mcmc_derive()
and mcmc_derive_data()
no longer vectorize new expression by defaultposterior_predictive_check()
now gives zeros by default.plot_data()
.pars.character()
internal only.Added ref_fun2
argument.
Added random_effects argument to predict etc to control random effects.
Merge pull request #30 from poissonconsulting/changes_to_actions.
Updating R CMD check Actions
replace chkor
Added zeros = FALSE argument to posterior_predictive_check() and type = NULL to residuals and simulate residuals.
Remove zeros from posterior predictive checks.
No longer export mcmc packages.
Merge branch 'master' of github.com:poissonconsulting/embr.
Renamed resample_residuals()
to simulate_residuals()
.
Added resample residuals which alters residual code!
posterior_predictive_check()
.fix up simplify = TRUE
add deltaIC
to glance()
add split and bound to rhat and converged
add deltaIC
to glance()
add marginalize_random_effects
argument to data_set
numericize difftime values as cause problems for STAN
set_analysis_mode()
now ensures all options are set unless mode = "reset"
and rhat
has been decreased from 1.1
to 1.05
for report mode and from 1.05
to 1.01
for paper mode.
parameters()
now returns character(0)
as opposed to NULL
if no parameters.
Rename derive()
and derive_data()
to mcmc_derive()
and mcmc_derive_data()
.
Moved pow()
to mcmcderive
package.
NEWS.md
file to track changes to the package.