I managed to solve the problem myself without using this code.
thx
2011-11-24 12:26 keltezéssel, Kehl Dániel írta:> Dear Community,
>
> I am trying to write code for the following problem.
> Lets assume we have a beta distribution.
> I know one quantile, lets say, 10% of the mass lies above .8, that is
> between .8 and 1.
> In addition, I know that the average of this "truncated tail" is
a
> given number, lets say .86.
> I have found the beta.select function in the LearnBayes package, which
> is as follows:
>
> function (quantile1, quantile2)
> {
> betaprior1 = function(K, x, p) {
> m.lo = 0
> m.hi = 1
> flag = 0
> while (flag == 0) {
> m0 = (m.lo + m.hi)/2
> p0 = pbeta(x, K * m0, K * (1 - m0))
> if (p0 < p)
> m.hi = m0
> else m.lo = m0
> if (abs(p0 - p) < 1e-04)
> flag = 1
> }
> return(m0)
> }
> p1 = quantile1$p
> x1 = quantile1$x
> p2 = quantile2$p
> x2 = quantile2$x
> logK = seq(-3, 8, length = 100)
> K = exp(logK)
> m = sapply(K, betaprior1, x1, p1)
> prob2 = pbeta(x2, K * m, K * (1 - m))
> ind = ((prob2 > 0) & (prob2 < 1))
> app = approx(prob2[ind], logK[ind], p2)
> K0 = exp(app$y)
> m0 = betaprior1(K0, x1, p1)
> return(round(K0 * c(m0, (1 - m0)), 2))
> }
>
> I assume one could change this code to get the results I need, but
> some parts of the function are not clear for me, any help would be
> greatly appreciated.
>
> Thanks a lot:
> Daniel
[[alternative HTML version deleted]]
