Prepare a Contingency Table for Attribute Assignment

Ensures that all conditioning-group combinations present in a synthetic population are represented in a contingency table.

Usage

prepare_contingency_table(
  contingency,
  synth_pop,
  group_by,
  target_attribute,
  strategy = "borrow"
)

Arguments

contingency

A contingency table containing the target attribute and a count column.

synth_pop

A synthetic population used to determine which conditioning-group combinations must be represented.

group_by

Character vector containing the conditioning variables used during attribute assignment.

target_attribute

Character string identifying the target attribute.

strategy

Character string specifying how missing contingency groups should be handled.

One of:

• "borrow"

• "overall"

• "error"

Value

A completed contingency table returned as a data.table.

Details

Missing contingency groups can be handled using one of three configurable strategies:

borrow

Borrow the nearest available conditional distribution.

overall

Use the overall target-attribute distribution.

error

Stop with an error if required groups are missing.

This function is typically invoked automatically by enhance before conditional attribute assignment begins.

The function compares all unique combinations of group_by variables found in the synthetic population against those present in the contingency table.

Any missing combinations are handled according to the specified strategy.

For "borrow", the function attempts to construct a distribution using a less-specific grouping level before falling back to the overall distribution.

For "overall", the function uses the overall target-attribute distribution computed from the contingency table.

For "error", an exception is raised whenever one or more required conditioning groups are missing.

This function prevents failures during attribute assignment caused by incomplete contingency tables and provides a configurable mechanism for handling sparse input data.

See also

ReplicaAdder, enhance, calculate_fractions, synthetic_population_to_contingency

Examples

if (FALSE) { # \dontrun{

population <- data.frame(
  age_group = c(
    "18-64",
    "18-64"
  ),
  gender = c(
    "Male",
    "Female"
  )
)

contingency <- data.frame(
  age_group = c(
    "18-64",
    "18-64",
    "18-64"
  ),
  gender = c(
    "Male",
    "Male",
    "Male"
  ),
  education = c(
    "Degree",
    "Diploma",
    "School"
  ),
  count = c(
    50,
    30,
    20
  )
)

expanded <- prepare_contingency_table(
  contingency = contingency,
  synth_pop = population,
  group_by = c(
    "age_group",
    "gender"
  ),
  target_attribute = "education",
  strategy = "borrow"
)

} # }