semsynth.metrics

Functions

heldout_loglik(model, df_test)
jensenshannon(p, q[, base, axis, keepdims])	Compute the Jensen-Shannon distance (metric) between two probability arrays.
js_divergence_discrete(p, q)
ks_2samp(data1, data2[, alternative, ...])	Performs the two-sample Kolmogorov-Smirnov test for goodness of fit.
per_variable_distances(real_df, synth_df, ...)
summarize_distance_metrics(distances)	Compute aggregate statistics for per-variable distance metrics.
wasserstein_distance(u_values, v_values[, ...])	Compute the Wasserstein-1 distance between two 1D discrete distributions.

semsynth.metrics.heldout_loglik(model, df_test: DataFrame) → Dict[str, float]

semsynth.metrics.js_divergence_discrete(p: Series, q: Series) → float

semsynth.metrics.per_variable_distances(real_df: DataFrame, synth_df: DataFrame, discrete_cols: List[str], continuous_cols: List[str]) → DataFrame

semsynth.metrics.summarize_distance_metrics(distances: DataFrame) → Dict[str, float]

Compute aggregate statistics for per-variable distance metrics.