skfolio.attribution.AssetBreakdown#

class skfolio.attribution.AssetBreakdown(names, vol_contrib, pct_total_variance, mu_contrib, weight, weight_std, systematic_vol_contrib, systematic_mu_contrib, idio_vol_contrib, idio_mu_contrib, vol, mu, corr_with_ptf)[source]#

Per-asset attribution breakdown.

Decomposes each asset’s volatility and return contribution into systematic and idiosyncratic components.

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

Attributes:

namesndarray of shape (n_assets,): Asset names. Always 1D.
weightndarray of shape (n_assets,) or (n_windows, n_assets): Portfolio asset weights.
weight_stdndarray of shape (n_assets,) or (n_windows, n_assets), or None: Standard deviation of asset weights over time. None when weights are not time-varying.
vol_contribndarray of shape (n_assets,) or (n_windows, n_assets): Total asset volatility contribution. Sums to total.vol.
systematic_vol_contribndarray of shape (n_assets,) or (n_windows, n_assets): Asset volatility contribution attributed to factor exposures. Sums to systematic.vol_contrib.
idio_vol_contribndarray of shape (n_assets,) or (n_windows, n_assets): Asset volatility contribution not attributed to factor exposures. Sums to idio.vol_contrib.
mu_contribndarray of shape (n_assets,) or (n_windows, n_assets): Total asset return contribution. Sums to total.mu.
systematic_mu_contribndarray of shape (n_assets,) or (n_windows, n_assets): Asset return contribution attributed to factor exposures. Sums to`systematic.mu`.
idio_mu_contribndarray of shape (n_assets,) or (n_windows, n_assets): Asset return contribution not attributed to factor exposures. Sums to idio.mu.
pct_total_variancendarray of shape (n_assets,) or (n_windows, n_assets): Percentage of total portfolio variance.
volndarray of shape (n_assets,) or (n_windows, n_assets): Standalone asset volatility: \(\sqrt{(B F B^\top + D)_{ii}}\).
mundarray of shape (n_assets,) or (n_windows, n_assets): Standalone asset return: expected return for predicted attribution and mean return for realized attribution.
corr_with_ptfndarray of shape (n_assets,) or (n_windows, n_assets): Asset correlation with portfolio returns.