R/MH_MCMC.R

In PratheepaJ/BARBI: A Bayesian Approach to Contamination Removal in Molecular Microbial Studies

#' MH_MCMC
#'
#' Metropolis Hastings MCMC
#'
#' @param itera numeric, number of MCMC samples.
#' @param startvalue_lamda_r numeric, initial value for the true intensity.
#' @inheritParams posterior
#'
#' @import HDInterval dplyr
#' @return An array of MCMC samples.
#' @export
MH_MCMC = function(itera, k,
                   al_c,
                   be_c,
                   startvalue_lamda_r = 0) {

    startvalue_lamda_r = startvalue_lamda_r
    chain_lamda_r = array(dim = c(itera + 1, 1))
    chain_lamda_r[1, ] = startvalue_lamda_r

    probab = numeric()

    # diff_shape_para_prop = k
    diff_shape_para_prop = (k - al_c)
    if (diff_shape_para_prop <= 0) {
        diff_shape_para_prop = (k + 0.5)
    }

    for (mc in 1:itera) {
        proposal = proposalFunction(al_p = diff_shape_para_prop,
                                    be_p = 1)

        pro = exp(posterior(lambda_r = proposal, k = k, al_c = al_c, be_c = be_c) -
            posterior(lambda_r = chain_lamda_r[mc, ], k = k, al_c = al_c,
                be_c = be_c))

        probab[mc] = min(c(1, pro))

        if (runif(1) < probab[mc]) {
            chain_lamda_r[mc + 1, ] = proposal
        } else {
            chain_lamda_r[mc + 1, ] = chain_lamda_r[mc, ]
        }
    }
    return(chain_lamda_r)
}