Currently mbest doesn't seem to be able to handle variables used for formulas (see the attempt to fit the model2 object below). If the formula isn't known until call time, one can use the do.call function as a work around. This presents another issue when one calls the summary method on the model object. In addition to the usual output, source code as well as a printout of any data.frame data argument are also printed out. For models fit on large datasets, this output can become overwhelming.
Either allowing for model formula variables, or resolving the summary call from models fit with do.call would be beneficial extensions.
Example
library(mbest)
print(version)
# platform x86_64-apple-darwin13.4.0
# arch x86_64
# os darwin13.4.0
# system x86_64, darwin13.4.0
# status
# major 3
# minor 3.2
# year 2016
# month 10
# day 31
# svn rev 71607
# language R
# version.string R version 3.3.2 (2016-10-31)
# nickname Sincere Pumpkin Patch
packageVersion("mbest")
# [1] ‘0.5’
sleepstudy <- lme4::sleepstudy
# fit model
model1 <- mhglm(Reaction ~ Days + (Days | Subject), gaussian, sleepstudy)
# fit model with formula variable
model_formula <- Reaction ~ Days + (Days | Subject)
model2 <- mhglm(model_formula, gaussian, sleepstudy)
# Error in mhglm(model_formula, gaussian, sleepstudy) :
# Invalid grouping factor specification, Subject
# In addition: Warning message:
# In Ops.factor(Days, Subject) : ‘|’ not meaningful for factors
# fit model with do.call
model_args <- list(
formula = Reaction ~ Days + (Days | Subject),
family = gaussian,
data = sleepstudy
)
model3 <- do.call(mhglm, model_args)
summary(model1)
summary(model3) # will be very long
Currently mbest doesn't seem to be able to handle variables used for formulas (see the attempt to fit the
model2
object below). If the formula isn't known until call time, one can use thedo.call
function as a work around. This presents another issue when one calls thesummary
method on the model object. In addition to the usual output, source code as well as a printout of any data.framedata
argument are also printed out. For models fit on large datasets, this output can become overwhelming.Either allowing for model formula variables, or resolving the summary call from models fit with
do.call
would be beneficial extensions.Example