smooth.OMG.confint

OMG.confint(parm=None, level=0.95, type=None, bootstrap=False, step_size=None, **boot_kwargs)

Confidence intervals for the joint OMG parameters.

Mirrors R’s confint.omg: SEs from the joint vcov(), t-quantile half-widths with R’s asymmetric degrees of freedom, then per-sub-model clamping to each side’s admissible region (via the sub-model’s own _clamp_confint_offsets()). Rows are prefixed A: / B: to identify which sub-model each parameter belongs to.

Parameters:

  • parm (str or sequence of str, optional) – Subset of (prefixed) names to return.

  • level (float) – Confidence level for the interval.

  • bootstrap (bool) – Not implemented for OMG yet.

  • step_size (float, optional) – Finite-difference step forwarded to vcov().

Returns:

Columns ["S.E.", "<lo>%", "<hi>%"] indexed by prefixed names.

Return type:

pandas.DataFrame

Parent Class: OMG