R/ffused.ada.R

In FastSF: Fast Structural Filtering

ffused.ada <- function (y, tau = 1, s.max = 20, eps = 0.1)
{
  n <- length(y)
  m = n-1

  kk <- 1
  beta.all <- NULL

  while (kk * tau < s.max) {
    s <- kk*tau
    re <- ffused(y = y, s = s, K.max = 5)
    beta <- re$beta
    beta.all <- cbind(beta.all, beta)
    z <- re$z
    u <- re$u
    mse <- mean((y - beta)^2)
    if (mse < eps | s > s.max) {
      break
    }
    else {
      kk <- kk + 1
    }
  }
  df <- 1:s
  return(list(y = y, beta = beta, v = z, beta.all = beta.all, df = df))
}