This helper function is used in glm_wrap to fit a list of models and select the best one. The model selection procedure will use the root mean square error (RMSE) metric implemented in yardstick::rmse to select the best model.
Arguments
- .data
Data including all explanatory and outcome variables needed for modeling; must include column for "location"
- .models
List of models defined as trending::trending_model objects
- complete
Logical as to whether or not all observations for covariates must be available in a given model; default is
TRUE
Value
A tibble containing characteristics from the "best" glm model including:
model_class: The "type" of model for the best fit
fit: The fitted model object for the best fit stored as a list column
location: The geographic unit being modeled
data: Original model fit data as a
tibblein a list column