pmse: Compute the pMSE metric between synthetic and real data

In synthesizer: Synthesize Data Based on Empirical Quantile Functions and Rank Order Matching

Compute the pMSE metric between synthetic and real data

Description

The propensity mean squared error is defined as \frac{1}{N}\sum_{i=1}^N(p_i-c)^2, where c is the number of synthetic records, divided by the sum of the number of synthetic and real records.

Usage

pmse(synth, real, model = c("lr", "rf"), nrep = NULL)

Arguments

synth	[data.frame] Synthesized data.
real	[real] Data to compare with the synthesized data.
model	[character] Model used to compute propensity scores. Options are "lr": logistic regression, and "rf": random forest.
nrep	[integer] Number of model repetitions to average the pMSE value over. Ignored for lr.

Value

[numeric] scalar.

Examples

scars <- synthesize(cars)
pmse(scars, cars)

synthesizer documentation built on April 12, 2025, 2:27 a.m.