quantlab.portfolio.predefined.ledoit_wolf#
A Ledoit-Wolf shrunk sample covariance of trailing one-bar returns.
LedoitWolfEstimator estimates the covariance of one-bar returns at a bar
from the lookback_bars returns ending there, shrunk toward a scaled
identity with the Ledoit-Wolf coefficient, so the estimate stays well
conditioned when there are more symbols than bars. The shrunk covariance is
split into correlations and volatilities, and the volatilities can be
replaced by given ones (a model’s forecast) before it is put back together.
It is kept in low-rank form, the scaled window plus a diagonal, never as the
dense [n, n] matrix (shrunk_low_rank, which the approximate
backtest’s batched estimate shares).
Classes
Ledoit-Wolf shrunk covariance of the trailing one-bar returns. |
Functions
Return |