R/mict.initial.R

In seqimpute: Imputation of Missing Data in Sequence Analysis

mict.initial <- function(data, imp, futureDistrib, REFORDI_L, MaxInitGapSize,
                         nr, nc, ud, nco, ncot, nfi, regr, k, available,
                         noise, ...) {
  OD <- data$OD
  covariates <- data$CO
  time.covariates <- data$COt
  COtsample <- data$COtsample

  if (nfi > nc - MaxInitGapSize) {
    nfi <- nc - MaxInitGapSize
  }

  train <- compute.traindata(data,
    MaxGap = 0, order = 0, shift = 0, np = 0, nc, nr,
    nfi, k, pastDistrib = FALSE, futureDistrib, col = 0,
    frame.radius = 0, regr, timing = FALSE
  )


  reglog <- fitmodel(train, regr, ...)

  for (order in 1:MaxInitGapSize) {
    imp <- impute(
      order = order, covariates, train,
      time.covariates, COtsample, OD, imp, pastDistrib = FALSE,
      futureDistrib, available,
      REFORDI_L, ncot, nc, np = 0, nfi, k, regr, reglog, noise,
      shift = 0, MaxGap = MaxInitGapSize
    )
  }
  return(imp)
}