fb4b9023f9ec516648d58b4ba2ecd8e241e21618,pyglmnet/utils.py,,tikhonov_from_prior,#Any#Any#,73

Before Change

    Given a prior covariance matrix, returns a Tikhonov matrix
    [U, S, V] = np.linalg.svd(PriorCov, full_matrices=False)
    Tau = np.dot(np.diag(1. / np.sqrt(S)), U)
    Tau = 1. / np.sqrt(np.float(n_samples)) * Tau / Tau.max()
    return Tau

After Change

    S_inv[zero_indices] = threshold
    S_inv[nonzero_indices] = 1. / S_inv[nonzero_indices]

    Tau = np.dot(np.diag(S_inv), V)
    n_features = Tau.shape[0]
    Tau = 1. / n_features * Tau
    return Tau

Instances