rankconf: Construct confidence intervals around ranks.

In harveybarnhard/rankconf: A Package for Rank Inference

Description

Construct confidence intervals around ranks.

Usage

rankconf(
  y,
  sig2,
  type = "PCER",
  method = "NAIVE",
  alpha = 0.05,
  k = 1,
  best = "min",
  thr = 2,
  nchains = 4,
  nwarmup = 1500,
  niter = 2500
)

Arguments

y	estimates
sig2	variance of estimates
type	What error should be controlled?
method	What method should be used to control error rates?
alpha	The rate at which the error should be controlled. 0.05 by default.
k	Number of errors to control, used for familywise error rates. 1 by default.
best	"min" for rank 1 to correspond to lowest estimate. "max" for rank 1 to correspond to greatest estimate.
thr	Number of threads to use for parallel processing. 2 by default.
nchains	For Bayesian Analysis: How many markov chains?
nwarmup	For Bayesian Analysis: How many warmup iterations?
niter	For Bayesian Analysis: How many total iterations?

harveybarnhard/rankconf documentation built on Jan. 2, 2022, 5:55 a.m.