yah

Author: microprediction

precise/covariance/statefunctions.py

@@ -6,36 +6,6 @@
 import numpy as np
 
 
-def ledoit_wolf()
-
-
-def naive_ledoit_wolf_shrinkage(n_samples, pcov, p2cov):
-    """
-    :param n_samples:
-    :param pcov:        Covariance of X
-    :param p2cov:       Covariance of
-    :return:
-    """
-    # From https://github.com/scikit-learn/scikit-learn/blob/main/sklearn/covariance/tests/test_covariance.py
-    # A simple implementation of the formulas from Ledoit & Wolf
-    # The computation below achieves the following computations of the
-    # "O. Ledoit and M. Wolf, A Well-Conditioned Estimator for
-    # Large-Dimensional Covariance Matrices"
-    # beta and delta are given in the beginning of section 3.2
-    n_features = np.shape(pcov)[0]
-    mu = np.trace(pcov) / n_features
-    delta_ = pcov.copy()
-    delta_.flat[:: n_features + 1] -= mu
-    delta = (delta_ ** 2).sum() / n_features
-    beta_ = 1.0 / (n_features * n_samples) * np.sum(p2cov - pcov ** 2)
-
-    beta = min(beta_, delta)
-    shrinkage = beta / delta
-    return shrinkage
-
-
-
-
 def oas(n_samples:int, pcov=None, scov=None, **ignore):
     """ Compute shrunk covariance matrix from empirical
     :param pcov:         Covariance matrix