neighbayes.diagnostics.SpatialCVResult

class neighbayes.diagnostics.SpatialCVResult(elpd, se, elpd_per_fold, n_per_fold, fold_ids, n_folds, method)[source]

Result of spatial_kfold().

elpd[source]

Expected log pointwise predictive density summed over folds.

Type:

float

se[source]

Standard error of elpd, estimated as sqrt(n * var(per_obs_elpd)) where per_obs_elpd spreads each fold’s elpd uniformly over its observations.

Type:

float

elpd_per_fold[source]

Per-fold elpd of shape (n_folds,).

Type:

np.ndarray

n_per_fold[source]

Number of observations in each fold, shape (n_folds,).

Type:

np.ndarray

fold_ids[source]

Fold in which each observation was held out, shape (n,); -1 for observations never held out. When a splitter tests an observation more than once, the last fold wins.

Type:

np.ndarray

n_folds[source]

Number of folds actually used.

Type:

int

method[source]

Class name of the splitter that produced the folds.

Type:

str

__init__(elpd, se, elpd_per_fold, n_per_fold, fold_ids, n_folds, method)[source]

Methods

__init__(elpd, se, elpd_per_fold, ...)

Attributes

elpd

se

elpd_per_fold

n_per_fold

fold_ids

n_folds

method

elpd[source]
elpd_per_fold[source]
fold_ids[source]
method[source]
n_folds[source]
n_per_fold[source]
se[source]