model_fit: Fit the insuRglm model

In realgabon/insuRglm: Tools for GLM Modeling in Insurance Context

Fit the insuRglm model

Description

Fits the model with the current model formula. Computes and saves back many new attributes and objects. This is a required step before using model_visualize, model_compare, model_save, model_betas, model_crossval, model_lift, model_export and others. In case of a big dataset (especially many columns), declaring future::plan(multiprocess) beforehand might help to speed up the process.

Usage

model_fit(setup, glm_backend = c("setup", "speedglm", "stats"))

Arguments

setup	Setup object. Created at the start of the workflow. Usually piped in from previous step.
glm_backend	Character scalar. Either 'setup', 'speedglm' or 'stats'. Choosing 'setup', which is a default choice results in using the glm_backend set during setup. Choosing 'speedglm' or 'stats' will temporarily override this by using speedglm::speedglm or stats::glm.

Value

Setup object with updated attributes.

Examples

require(dplyr) # for the pipe operator
data('sev_train')

setup <- setup(
  data_train = sev_train,
  target = 'sev',
  weight = 'numclaims',
  family = 'gamma',
  keep_cols = c('pol_nbr', 'exposure', 'premium')
)

# parallel processing is supported and may be faster on bigger datasets
plan(multiprocess)

modeling <- setup %>%
  factor_add(pol_yr) %>%
  factor_add(agecat) %>%
  model_fit()

modeling %>%
  model_visualize(factors = 'fitted')

modeling %>%
  model_visualize(factors = 'unfitted')

realgabon/insuRglm documentation built on Jan. 2, 2023, 2:51 a.m.