Add the capacity to deal with random structure

[beckyfisher]

No description provided.

[beckyfisher]

This capacity is now implemented in the experimental branch random_effects. This is achieved through two additional arguments. The first, random, is a named list containing model formula for the random effects that you want to include. This can include all/any of the non-liner parameters, and bayesnec will automatically use the ones relevant to a given model. The second is a character vector specifying the columns in data that contain the random effects. Below is an example of the code for implementing the random effects, where "genotype" is added as a random effect on the "nec" parameter and "chamber" is added as an offset term (ost). Note that for random offsets, the syntax requires the random list includes this as element "ost", and that "ost" is used in the model formula. Note also that this example is likely to be a sensible model, and that use of the random effects structures needs to be carefully considered, explored and validated by the user.

random_term <- list( nec = as.formula(nec ~ (1|genotype)),
ost = as.formula(ost ~ (1|chamber)))

temp5<- bnec(data = surv_24h_data,
x_var = "log.x",
y_var = "y",
trials_var = "trials",
model = "ecxll4",
iter=2e3,
random = random_term, random_vars = c("genotype", "chamber")

[beckyfisher]

Note this has been re-written in version 2.0 such that the random structure is closer to the brms syntax.
Still requires vignette to demonstrate usage.