Skip to contents

  • Impute
  • One function per method, never one function with a method= argument. Single mean imputation is the SAS imputsub / PROC STANDARD REPLACE job; multiple imputation (PROC MI / mult_imput) is a prototype, not yet hardened or released, and does not pool its m completed datasets.

    impute_mean()
    Fill missing values with the variable mean
    impute_multiple()
    Fill missing values with multiple imputation

  • Read the record
  • Every imputation returns the completed data and a record of exactly what was changed. The matrix is the primary artifact; everything else here is a view of it.

    hvti_imputation
    The imputation record
    hvti_imputation_multi
    The multiple-imputation record
    imputed_data() imputed_matrix() imputed_any() complete_case_pass() analysis_pass() kept_by_imputation() imputation_provenance()
    Read an imputation record
    imputation_indicators()
    Missingness indicators, generated from the record
    summary(<hvti_imputation>)
    Summarise what was imputed
    summary(<hvti_imputation_multi>)
    Summarise what was imputed, across the m completed datasets

    Package

    hvtiRimputation hvtiRimputation-package
    hvtiRimputation: Missing Value Imputation for the HVTI CORR Group