testassay (version 0.1.0)

normConstCVCI: Log-centered confidence intervals from a Normal constant coeffficient of variation model

Description

Assume Y is normal with mean mu>0 and coefficient of variation theta, then Y/mu ~ N(1, theta^2). Get log-centered confidence intervals (when possible), meaning intervals such that log(y) +/- r(theta), where r(theta) is a constant function of theta.

Usage

normConstCVCI(y, theta, conf.level = 0.6827, eps = .Machine$double.eps^0.25)

Arguments

y

vector of observed values, should be positive

theta

coefficient of variation (assumed known)

conf.level

Confidence level

eps

a small number used in the algorithm (look at code before changing)

Value

A list with the following components

• obs y

• lower lower confidence limit on mu=E(Y)

• upper upper confidence limit on mu=E(Y)

Examples

# NOT RUN {
# defaults to 68.27 percent confidence level, same level as Normal plus or minus 1 std dev.
normConstCVCI(3.4,.6)
# symmetric on log scale
log(normConstCVCI(3.4,.6))

# }