bayesPoisTest: Bayesian two-sample test of Poisson rates

In dkahle/bayesRates: Two-Sample Tests and Sample Size Determination from a Bayesian Perspective

Description

Bayesian two-sample test of Poisson rates

Usage

bayesPoisTest(x, t, a1, b1, a2, b2, a = a1, b = b1, pi0 = 0.5,
  pi1 = 1 - pi0, c = 1, rule = c * pi1/pi0)

Arguments

x	two-vector containing occurrences
t	duration of trials (the same for both samples)
a1	alpha, the hyperparameter of the gamma distribution of the first poisson rate
b1	beta, the hyperparameter of the gamma distribution of the first poisson rate
a2	alpha, the hyperparameter of the gamma distribution of the second poisson rate
b2	beta, the hyperparameter of the gamma distribution of the second poisson rate
a	alpha, the hyperparameter of the gamma distribution under the null
b	beta, the hyperparameter of the gamma distribution under the null
pi0	the prior probability of the null hypothesis
pi1	the prior probability of the alternative hypothesis
c	relative loss constant (loss due to type II error divided by loss due to type I error)
rule	the decision rule for the Bayes factor B. by default, test rejects when B < pi1/pi0 * c2/c1

See Also

samplePower, samplePowerEst, findSize

Examples

## Not run: 

a1 <- 5; b1 <- .2 # E = 1
a2 <- 30; b2 <- .167 # E = 5
plotGamma(c(a1, a2), c(b1, b2))
bayesPoisTest(c(18, 105), 20, a1, b1, a2, b2)
bayesPoisTest(c(18, 165), 20, a1, b1, a2, b2)

# compare to :
prop.test(c(10, 16), c(20, 20))


 
## End(Not run)

dkahle/bayesRates documentation built on May 15, 2019, 9:07 a.m.