R/models__linear__etc__idx.beta2b.R

In feng-li/movingknots: Efficient Bayesian multivariate surface regression

#' Convert indices from beta to b
#'
#' Details from the paper
#' @param p NA
#' @param q NA
#' @param q.i NA
#' @return NA
#' @author Feng Li, Department of Statistics, Stockholm University, Sweden.
#' @export
idx.beta2b <- function(p, q, q.i)
  {
    idx4beta <- matrix(1:(p*q), q)
    cum.q.i <- c(0, cumsum(q.i))

    out4b <- matrix(0, 1:(p*q))
    for(i in 1:length(q.i))
      {
        idx4bi <- (1 + cum.q.i[i]):cum.q.i[i+1]
        out4b[(1 + cum.q.i[i]*p):(cum.q.i[i+1]*p)] <- idx4beta[idx4bi, ]
      }
    return(out4b)
  }