What leaves the study: the fingerprint of a fitted model

The part of a fitted model that synthetic data is generated from, and nothing else. synpmx_model_estimate() returns a fit that also carries diagnostics for the person who ran it: the candidate comparison and its AIC values, the starting values, the correlations between covariates and the individual random effects, the design notes and timings. None of those is needed to generate, so none of them is in the fingerprint. synpmx_model_generate() reduces whatever it is given to its fingerprint before generating, and attaches the fingerprint, never the fit, to the data it returns.

Usage

model_fingerprint(fitted_model)

model_release(fitted_model)

Arguments

fitted_model

A pmx_fitted_model from synpmx_model_estimate(), or a fingerprint, which is returned unchanged.

Value

A pmx_model_fingerprint, which is also a pmx_fitted_model.

Details

Save the fingerprint, not the fit, when it has to be carried out of the environment that holds the study: synpmx_model_generate() accepts either and produces the same data from both. model_release() is another name for the same function.

See also

model_privacy_checks(), synpmx_model_generate(), model_report().