arviz_stats.ELPDDataLOOKFold

class arviz_stats.ELPDDataLOOKFold(*, elpd, se, p, n_samples, n_data_points, scale, warning, good_k, elpd_i=None, pareto_k=None, n_folds, p_kfold_i=None, fold_fits=None)

K-fold cross-validation results returned by loo_kfold.

Inherits the attributes of ELPDData. good_k and pareto_k are always None because k-fold cross-validation does not use importance sampling.

Attributes:

n_foldsint

Number of folds.

p_kfold_ixarray.DataArray, optional

Pointwise effective number of parameters, only when pointwise=True.

fold_fitsdict, optional

Fitted models for each fold, only when save_fits=True.

__init__(*, elpd, se, p, n_samples, n_data_points, scale, warning, good_k, elpd_i=None, pareto_k=None, n_folds, p_kfold_i=None, fold_fits=None)

Methods

__init__(*, elpd, se, p, n_samples, ...[, ...])

Attributes

elpd_i
fold_fits
kind	Name of the estimator that produced the result.
p_kfold_i
pareto_k
n_folds
elpd
se
p
n_samples
n_data_points
scale
warning
good_k