skfolio.attribution.FamilyBreakdown#

class skfolio.attribution.FamilyBreakdown(names, vol_contrib, pct_total_variance, mu_contrib, exposure, exposure_std, mu_contrib_uncertainty=None)[source]#

Family-level attribution breakdown.

Aggregates factor attribution by factor family.

For single-point attribution, arrays have shape (n_families,). For rolling attribution, arrays have shape (n_windows, n_families).

Attributes:

namesndarray of shape (n_families,): Family names. Always 1D.
exposurendarray of shape (n_families,) or (n_windows, n_families): Sum of portfolio factor exposures within each family.
exposure_stdndarray or None: Standard deviation of family exposures over time. None for predicted attribution.
vol_contribndarray of shape (n_families,) or (n_windows, n_families): Family volatility contribution, equal to the sum of its factor volatility contributions.
pct_total_variancendarray of shape (n_families,) or (n_windows, n_families): Percentage of total portfolio variance.
mu_contribndarray of shape (n_families,) or (n_windows, n_families): Family return contribution, equal to the sum of its factor return contributions.
mu_contrib_uncertaintyndarray of shape (n_families,) or (n_windows, n_families) or None: Standard error of the family mean return contribution, accounting for cross-factor estimation correlations within the family. None when uncertainty is not computed.