numeraire.baselines.MeanVariance#

class numeraire.baselines.MeanVariance(*, normalization: Literal['budget', 'none'] = 'budget', window: int | None = None, min_obs: int | None = None)[source]#

Bases: object

Plug-in mean-variance estimator: sample mu/S → S^-1 mu, normalization explicit.

Parameters:

  • normalization – "budget" (default) divides by 1' S^-1 mu (weights sum to one; DGU convention); "none" returns the raw proportional direction S^-1 mu.

  • window – As for MinVariance (rolling estimation window / warm-up; default warm-up is one row more than the asset count so the sample covariance is invertible).

  • min_obs – As for MinVariance (rolling estimation window / warm-up; default warm-up is one row more than the asset count so the sample covariance is invertible).

__init__(*, normalization: Literal['budget', 'none'] = 'budget', window: int | None = None, min_obs: int | None = None) → None[source]#

Methods

__init__(*[, normalization, window, min_obs])

fit(view)