Description Usage Arguments Note References See Also Examples
Compute the power of the one-sample equivalence test for population means of normal summary values with known population variance.
1 2 |
rho |
vector of quantile |
tau.u |
Upper boundary point of the equivalence region |
alpha |
Level of the equivalence test |
sigma |
Standard deviation of the test statistic. |
norm |
Normalization constant for the truncated power function. |
support |
Support of the truncated power function (vector of dimension 2). |
log |
If |
The power function can be truncated to support
and then standardized with norm
.
If one of these is set, the other must be provided too.
http://arxiv.org/abs/1305.4283
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 | # power function of the Z-test,
# to test the equivalence of autocorrelations of normal summary values in the MA(1) model
n2s <- function(n){ 1/sqrt(floor(n)-3) } #need formula to convert n.of.y into s.of.T
s2n <- function(s){ (1/s)^2+3 } #need formula to convert s.of.T into n.of.y
n.of.y <- 1500 #number of independent pairs of the form (y_t, y_{t+1}) for simulated summary values y_t, t=1, ...
# reasonable power properties of the ABC accept/reject step
rho <- seq(-0.3,0.3,0.001)
tau.u <- 0.09
tmp <- data.table(rho=rho, power=ztest.pow(rho, tau.u, alpha=0.01, n2s(n.of.y)))
p <- ggplot(tmp,aes(x=rho,y=power)) + geom_line() + labs(y='Power\n(ABC acceptance probability)')
print(p)
# power increases with tau.u and becomes flat
tmp <- lapply(c(0.06,0.07,0.08,0.1,0.15,0.2),function(tau.u)
{
data.table(tau.u=as.factor(tau.u),rho=rho,power=ztest.pow(rho, tau.u, alpha=0.01, n2s(n.of.y)), tau.u=tau.u)
})
tmp <- do.call('rbind', tmp)
p <- ggplot(tmp,aes(x=rho,y=power,colour=tau.u,group=tau.u)) + geom_line() + labs(y='Power\n(ABC acceptance probability)')
print(p)
# power increases with n.of.y and becomes flat
tau.u <- 0.1
tmp <- lapply(c(750,1000,1500,3000),function(n.of.y)
{
data.table(n.of.y=as.factor(n.of.y),rho=rho,power=ztest.pow(rho, tau.u, alpha=0.01, n2s(n.of.y)),n.of.y=n.of.y)
})
tmp <- do.call('rbind', tmp)
p <- ggplot(tmp,aes(x=rho,y=power,colour=n.of.y,group=n.of.y)) + geom_line() + labs(y='Power\n(ABC acceptance probability)')
print(p)
|
Add the following code to your website.
For more information on customizing the embed code, read Embedding Snippets.