NEWS.md

syntheval 0.0.4

• Add empirical disclosure risk metrics.
• Add comparison visualization utilities.
• Add na.rm functionality to most functions to handle NA values.
• Add families to roxygen2 headers.
• Ensure that all functions return ungrouped output.

syntheval 0.0.3

• Add a README with examples.
• Add flexible system for calculating discriminant-based metrics for utility including the pMSE, pMSE ratio, SPECKS, and discriminator AUC. The code can use most functionality from library(tidymodels) to generate propensities.
• Add common_vars and synth_vars arguments to most utility functions. These control the variables included in the utility metrics.

syntheval 0.0.2

New Features

• Add util_co_occurrence() for comparing co-occurrence matrices from the original and synthetic data.
• Add util_ks_distance() for calculating the KS distance (D) for each numeric variable in the original and synthetic data.

Improvements

• Use reframe() in util_percentiles() instead of summarize() to minimize warnings.

syntheval 0.0.1

New Features

• Create first numbered version of library(syntheval)!
• Add util_proportions(), util_totals(), util_moments(), and util_percentiles() for comparing univariate statistics in the original and synthetic data.
• Add util_corr_fit() for comparing the correlation matrices from the original and synthetic data.
• Add util_ci_overlap() for calculating regression confidence interval overlap.
• Add pmse(), null_pmse(), and pmse_ratio() for calculating one version of the pMSE ratio.
• Add disc_mit() for a basic membership inference test.