PTMLocScalePredictions

class liesel_ptm.PTMLocScalePredictions(samples, model, y=None, **kwargs)

Bases: object

Posterior predictions for a penalized transformation model.

Parameters:

• samples (dict[str, Any]) – Dictionary of posterior samples.

• model (PTMLocScale) – The model object.

• y (Optional[Any]) – Response observations to use in predictions. If None``(default), the observed values are extracted from the model object. (default: ``None)

• **kwargs – Values to use for all terms in the location and scale model parts. Must have appropriate shapes. For terms that are not explicitly specified, the observed values are extracted from the model object.

Methods

predict_cdf	Predicts the response's conditional cumulative distribution function.
predict_loc	Predicts the location.
predict_log_prob	Predicts the response's conditional log probability.
predict_normalization_and_deriv	Predicts the transformed residuals and the derivative of the normalization.
predict_normalization_deriv	Predicts the derivative of the normalization.
predict_normalization_inverse	Predicts the inverse of the normalization function.
predict_pdf	Predicts the response's conditional probability density.
predict_quantile	Predicts the conditional response quantiles at probability level p.
predict_resid_quantile	Predicts the residual quantiles at probability level p.
predict_residual_cdf	Evaluates the residual's posterior cumulative distribution function.
predict_residual_log_prob	Evaluates the residual's posterior log probability function.
predict_residual_pdf	Evaluates the residual's posterior density function.
predict_residuals	Predicts the residuals.
predict_scale	Predicts the location.
predict_transformation	Predicts the transformed residuals.
predict_transformation_deriv	Predicts the derivative of the transformation.
predict_transformation_inverse	Predicts the inverse of the normalization function.
predict_z	Predicts the transformed residuals.
predict_z_deriv	Predicts the derivative of the transformation.
sample	Generates random samples from the posterior predictive distribution.
summarise_density_by_quantiles	Returns a summary dataframe to quickly assess the response's conditional distribution.
summarise_density_by_samples	Returns a summary dataframe to quickly assess the response's conditional distribution based on a random sample from the posterior.
summarise_transformation_by_quantiles	Returns a summary dataframe to quickly assess the transformation function.

Attributes

samples	The posterior samples.
model	The model object.