conformal_inf_linear: Conformal inference procedure to compute a confidence...

View source: R/inference.R

conformal_inf_linearR Documentation

Conformal inference procedure to compute a confidence interval for a linear in time effect

Description

Conformal inference procedure to compute a confidence interval for a linear in time effect

Usage

conformal_inf_linear(
  ascm,
  alpha = 0.05,
  stat_func = NULL,
  type = "iid",
  q = 1,
  ns = 1000,
  grid_size = 50
)

Arguments

ascm

Fitted 'augsynth' object

alpha

Confidence level

stat_func

Function to compute test statistic

type

Either "iid" for iid permutations or "block" for moving block permutations; default is "iid"

q

The norm for the test static '((sum(x ^ q))) ^ (1/q)'

ns

Number of resamples for "iid" permutations

grid_size

Number of grid points to use when inverting the hypothesis test

Value

List that contains:

  • "att"Vector of ATT estimates

  • "heldout_att"Vector of ATT estimates with the time period held out

  • "se"Standard error, always NA but returned for compatibility

  • "lb"Lower bound of 1 - alpha confidence interval

  • "ub"Upper bound of 1 - alpha confidence interval

  • "p_val"p-value for test of no post-treatment effect

  • "alpha"Level of confidence interval


ebenmichael/augsynth documentation built on March 20, 2024, 5:20 a.m.