Relative change in 95 % HPDI width between the last third and the full post-warmup chain, in percent. Large drift indicates the chain has not yet visited the tails enough times for the interval estimate to stabilise.

Type:¶: dict[str, float]

warnings_triggered[source]¶

Human-readable description of each adequacy check that failed.

Type:¶: list of str

adequate[source]¶

True when no warnings were triggered for the spatial dependence parameter(s).

Type:¶: bool

adequate_by_param[source]¶

Per-parameter adequacy flag. True for a parameter when none of the adequacy checks (ESS, yield, HPDI drift, R-hat) failed for it. Only populated for the auto-detected spatial dependence parameters; extra_params are not gated.

Type:¶: dict[str, bool]

__init__(parameters, ess_bulk, ess_tail, r_hat, mcse_mean, nominal_size, yield_pct, hpdi_drift_pct, warnings_triggered=<factory>, adequate=True, adequate_by_param=<factory>)[source]¶

Methods

`__init__`(parameters, ess_bulk, ess_tail, ...)
`to_frame`()	Return per-parameter diagnostics as a `pandas.DataFrame`.

Attributes

`adequate`
`parameters`
`ess_bulk`
`ess_tail`
`r_hat`
`mcse_mean`
`nominal_size`
`yield_pct`
`hpdi_drift_pct`
`warnings_triggered`
`adequate_by_param`

adequate = True[source]¶

adequate_by_param[source]¶

ess_bulk[source]¶

ess_tail[source]¶

hpdi_drift_pct[source]¶

mcse_mean[source]¶

nominal_size[source]¶

parameters[source]¶

r_hat[source]¶

to_frame()[source]¶

Return per-parameter diagnostics as a pandas.DataFrame.

One row per spatial dependence parameter (self.parameters), with columns for each per-parameter metric plus an adequate column derived from adequate_by_param. Global scalars (nominal_size) and warnings_triggered are not included; access them on the dataclass directly.

Returns:¶: Index name "parameter"; columns ordered per _FRAME_COLUMNS.
Return type:¶: pandas.DataFrame

warnings_triggered[source]¶

yield_pct[source]¶