Estimation of sum of model output

Calculate the variance between a baseline sum of predictions against a counterfactual sum of predictions in a gam based on modified code from gam predict function help

Usage

calc_sum_counterfactual_gam(
  m,
  baseline_data,
  counter_data = NULL,
  ci = 0.95,
  use_relative_diff = FALSE,
  method = "delta",
  nrep = 1000
)

Arguments

m

model object

baseline_data

data.frame of baseline predictors

counter_data

data.frame of counterfactual predictors. Optional can be NULL which would then just calculate variance of baseline sum.

ci

range of confidence interval

use_relative_diff

provide estimates as a relative difference, otherwise presented as an absolute difference

method

Selects which method is used to estimate confidence intervals and expectation. Can be one of the following:

• "delta"Implements delta approximation.

• "posterior"Implements posterior sampling. Use nrep to specify number of bootstrap samples

• "bootstrap"Implements simple bootstrap. Use nrep to specify number of bootstrap samples

• "efron bootstrap"Implements Efron bootstrap. Use nrep to specify number of bootstrap samples

nrep

number of samples used for posterior sampling. Only used if using posterior or bootstrap methods is TRUE

Value

list