bss.dt.lambarDP: Bayesian sample size in a decision-theoretic approach for the...

In eliardocosta/ssdet: Sample size determination in frequentist and Bayesian approaches

Description

Bayesian sample size in a decision-theoretic approach for the functional mean of the Dirichlet process with a gamma distribution as the $F_0$ base distribution.

Usage

bss.dt.lambarDP(lf, alpha, lam0, theta0, w, c, rho = NULL, gam = NULL,
  nmax = 100, nlag = 10, nrep = 10, lrep = 50, plot = FALSE,
  ncore = NULL, ...)

Arguments

lf	1 or 2, representing the loss function used.
alpha	Shape parameter of the Dirichlet process.
lam0	A positive real number representing a hyperparameter of the $F_0$ base distribution.
theta0	A positive real number representing a hyperparameter of the $F_0$ base distribution. We consider $F_0$ as the gamma distribution with mean $lam_0$ and shape parameter $theta_0$.
w	A positive real number representing the aliquot volume.
c	A positive real number representing the cost of colect one aliquot.
rho	A number in (0, 1). The probability of the credible interval is $1-rho$. Only for lost function 1.
gam	A positive real number connected with the credible interval when using lost function 2.
nmax	A positive integer representing the maximum number for compute the Bayes risk. Default is 100.
nlag	A positive integer representing the lag in the n's used to compute the Bayes risk. Default is 10.
nrep	A positive integer representing the number of samples taken for each $n$.
lrep	A positive integer representing the number of samples taken for $S_n$. Default is 50.
plot	Boolean. If TRUE (default) it plot the estimated Bayes risks and the fitted curve.
ncore	Number of cores to use in parallel computin. If NULL the function uses 1 core if there is only one core, if there is more than one cores uses one half of the cores.
...	Currently ignored.

Value

An integer representing the sample size.

eliardocosta/ssdet documentation built on Dec. 14, 2021, 6:27 a.m.