elicit.targets module#
- elicit.targets.computation_elicited_statistics(target_quantities: Dict[str, Tensor], targets: List[Target]) Dict[str, Tensor][source]#
Computes the elicited statistics from the target quantities by applying a prespecified elicitation technique.
- Parameters:
- target_quantitiesdict[str, tf.Tensor], shape: [B,num_samples,num_obs]
simulated target quantities.
- targetslist[dict]
list of target quantities specified with
elicit.elicit.target().
- Returns:
- elicits_resdict[res, tf.Tensor], shape: [B, num_stats]
simulated elicited statistics.
- elicit.targets.computation_target_quantities(model_simulations: Dict[str, Tensor], prior_samples: Tensor, targets: List[Target]) Dict[str, Tensor][source]#
Computes target quantities from model simulations.
- Parameters:
- model_simulationsdict[str, tf.Tensor]
simulations from generative model.
- prior_samplestf.Tensor; shape = [B, rep, num_params]
samples from prior distributions of model parameters. Currently only needed if correlations between model parameters is used as elicitation technique.
- targetslist[dict]
list of target quantities specified with
elicit.elicit.target().
- Returns:
- targets_resdict[str, tf.Tensor]
computed target quantities.