imputation_stats Objects

The imputation_stats class stores detailed information about a missing data imputation operation performed by impute_obs. These objects provide comprehensive statistics about what was imputed, what was skipped, and why.

Object Structure

An imputation_stats object is a list with class "imputation_stats" containing:

method: Character string indicating the imputation method used ("mean" or "median").
min_nonmiss_prop: Numeric value (0 to 1) indicating the minimum proportion of non-missing data required for a probe to be imputed.
timestamp: POSIXct timestamp of when imputation was performed.
n_total_probes: Total number of probes in the methylation matrix.
n_complete_probes: Number of probes with no missing values.
n_completely_missing_probes: Number of probes missing in all samples.
n_partially_missing_probes: Number of probes with some (but not all) missing values.
n_probes_imputed: Number of probes that had values imputed.
n_probes_skipped: Number of partially missing probes that were skipped because they didn't meet the min_nonmiss_prop threshold.
n_values_imputed: Total number of individual missing values that were imputed.
n_missing_before_imputation: Total missing values before imputation.
n_missing_after_imputation: Total missing values remaining after imputation.
imputation_rate: Proportion of originally missing values that were imputed (0 to 1).
probes_imputed: Character vector of CpG probe IDs that were imputed.
probes_skipped: Character vector of CpG probe IDs that were skipped.
values_imputed_per_probe: Named numeric vector showing the number of values imputed for each probe (only present when return_stats = TRUE).
skipped_reasons: Named character vector explaining why each skipped probe was not imputed (only present when return_stats = TRUE).

Creating imputation_stats Objects

Objects are created automatically when calling impute_obs with return_stats = TRUE:


# Perform imputation with statistics
result <- impute_obs(
  my_surro,
  method = "mean",
  min_nonmiss_prop = 0.5,
  return_stats = TRUE
)

# Access the statistics
stats <- result$imputation_stats

Accessing Information

Components can be accessed using standard list notation:


# View imputation summary
stats$n_values_imputed
stats$imputation_rate

# View which probes were imputed
stats$probes_imputed

# View detailed reasons for skipped probes
stats$skipped_reasons

Print Method

The package provides a print method that displays a formatted summary:


print(stats)

This shows an organized view of the imputation operation and its results.

Object Structure

Creating imputation_stats Objects

Accessing Information

Print Method

See also