Correctness of boundsPostprob() in parE and weights in beta Mixture cases

[audreyyeoCH]

Some things to check:

1. reverse engineer this function with postprob() as a test and see if it passes
2. this may be wrong to t() the parEtwice, one outside the helper postprob and one within. It may be more correct to cbind(c(alpha_1, beta_1), c(alpha_2, beta_2)) hence point 1 to reverse engineer in a test case, after all h_getbetaMixPost() calculated updated weights and parameters with :
   lbeta(a = postPar[, 1], b = postPar[, 2]) - lbeta(a = par[, 1], b = par[, 2]) therefore the parE has to adjust accordingly
3. if 1,2, will indeed be actionables, then fix user facing function, examples and test case.