Predict methods in ipmr

This function is used when a predict method is incorporated into the vital rate expressions of a kernel. Generally, ipmr can handle this without any additional user effort, but some model classes will fail (often with an obscure error message). When this happens, use_vr_model can ensure that model object is correctly represented in the data_list.

use_vr_model(model)

Arguments

model

A fitted model object representing a vital rate. Primarily used to avoid writing the mathematical expression for a vital rate, and using a predict() method instead.

Value

A model object with a "flat_protect" attribute.

Details

ipmr usually recognizes model objects passed into the data_list argument automatically. Unfortunately, sometimes it'll miss one, and the user will need to manually protect it from the standard build process. This function provides a wrapper around that process. Additionally, please file a bug report here: https://github.com/padrinoDB/ipmr/issues describing what type of model you are trying to use so it can be added to later versions of the package.

Wrap a model object in use_vr_model when building the data_list to pass to define_kernel.

Examples

data(iceplant_ex)

grow_mod <- lm(log_size_next ~ log_size, data = iceplant_ex)
surv_mod <- glm(survival ~ log_size, data = iceplant_ex, family = binomial())

data_list <- list(
  grow_mod = use_vr_model(grow_mod),
  surv_mod = use_vr_model(surv_mod),
  recruit_mean = 20,
  recruit_sd   = 5
)