PropensityScore.fit#

PropensityScore.fit(X, t, coords, prior=None, noncentred=True)[source]#

Draw samples from posterior, prior predictive, and posterior predictive distributions. We overwrite the base method because the base method assumes a variable y and we use t to indicate the treatment variable here.

Parameters:

  • X (ndarray) – Covariate matrix used to predict the treatment.

  • t (ndarray) – Observed treatment indicator (0/1).

  • coords (dict[str, Any]) – Coordinate names for named dimensions of the model.

  • prior (dict[str, list] | None) – Prior specification overrides. Defaults to {"b": [0, 1]}.

  • noncentred (bool) – Forwarded to build_model().

Return type:

InferenceData