get_avg_width: Get the empirical width of a confidence sequence

In WannabeSmith/drconfseq: Confidence sequences for the average treatment effect

Get the empirical width of a confidence sequence

Description

Get the empirical width of a confidence sequence

Usage

get_avg_width(data_generator_fn, conf_set_fn, times, num_repeats, n_cores = 1)

Arguments

data_generator_fn	A function which generates a vector of data (() -> real)
conf_set_fn	A function which takes a vector of data and produces a sequence of confidence intervals or a confidence sequence (real -> (real, real))
times	The times for which the miscoverage should be considered. For example, we might want to start the experiment late since early start times can lead to poor asymptotic approximations (a vector of positive integers)
num_repeats	How many repeats to perform. Higher values lead to better estimates of the miscoverage rate (positive int).
n_cores	How many cores to parallel process on (positive int).

Value

A vector of (increasing) miscoverage rates at each time (a (0, 1)-valued vector).

WannabeSmith/drconfseq documentation built on Sept. 13, 2023, 3:05 a.m.