noNegsneedMargin: Raking Algorithm Function C: when negative values are NOT...

In bcgov/dbutils: Utility functions to process and modify BC Stats' population databases

Raking Algorithm Function C: when negative values are NOT allowed and margin IS needed

Description

When raking, the appropriate algorithm is chosen (negative values allowed or not, margin needed or not), "whole people" adjustments to be made to selected cells in a row are calculated, then "residual people" adjustments to be made to selected cells in a row are calculated, then column control totals are reconciled. Raking is run iteratively row-by-row. Regardless of the algorithm needed, five arguments are needed. A sixth argument, "needMargin" is only needed for algorithm C (noNegsneedMargin). This is a helper function used in dbRake when negative values are NOT allowed AND margin IS needed. Specifically, values need to be added to cells in current row, and subtracted from lower rows. Because adjustments are subtracted from lower rows but negative values are not allowed, a margin is needed.

Usage

noNegsneedMargin(
  CurrRow,
  CurrRow_value,
  data,
  n_colGrps,
  n_rowGrps,
  RowAdj,
  needMargin
)

Arguments

CurrRow	a subset of data with 3 rows and columns: XXX, n_colGrps, Sum, Ctrl_TOTAL, Diff, adj_value, where XXX is Sex in Part 1, Region in Part 2, and Age in Part 3. The rows are the one XXX row currently being worked on, Adjustments, and AdjCurrRow.
CurrRow_value	a counter to iterate through rows in data
data	a dataframe of rows to be iteratively raked with columns: XXX, n_colGrps, Sum, Ctrl_TOTAL, Diff, adj_value, where XXX is the variable being raked (Sex in Part 1, Region in Part 2, Age in Part 3). Rows are the rows of XXX as well as three summary rows: Sum, Ctrl_TOTAL, Diff.
n_colGrps	the number of column groups to rake over (e.g., 89 LHAs, number of Age Groups, etc.)
n_rowGrps	the number of rows groups to rake over (e.g., 2 Sexes, 89 LHAs, etc.)
RowAdj	vector of zeros, of length n_colGrps (e.g., 89 zeros, etc.)
needMargin	a logical (TRUE or FALSE) of whether a margin is needed. The margin is the current row's Sum - CurrRow values - Adjustments - all rows above.

Value

original dataframe, but with CurrRow's Diff now 0, and any adjustments made to CurrRow reversed in the row below that has the largest minimum value

See Also

The overall raking function: dbRake()

Other raking helpers: add.random.fraction.to.cols(), adjustSex(), allowNegsnoMargin(), calc.cols(), multiRake(), noNegsnoMargin(), prep.prorate.col(), prorate.col(), prorate.row(), prorate(), read.inputs(), real.to.int(), rename.age.grps(), rounded()

bcgov/dbutils documentation built on Sept. 30, 2022, 12:04 a.m.