huge_glasso

Usage

huge_glasso(
    x,
    lambda_=None,
    nlambda=None,
    lambda_min_ratio=None,
    scr=None,
    cov_output=False,
    verbose=True,
    backend="native",
    *,
    input_type="auto",
) -> HugeResult

Description

Convenience wrapper for huge(..., method="glasso"). Use input_type="data" for square symmetric observations or input_type="covariance" to require covariance/correlation input. With no explicit lambda_, auto-detected covariance input matches R's historical diagonal-sensitive lambda scale; explicit covariance routing uses only off-diagonal entries. An indefinite pairwise covariance estimate is accepted only when regularization produces a finite positive-definite precision estimate and a certified covariance/precision pair.

An explicit lambda_ must be a numeric scalar, NumPy 0-D value, or non-empty one-dimensional sequence; multidimensional inputs are rejected. Values must be finite, strictly positive, and non-increasing; tied values are allowed.

Notes

• cov_output=True is supported for glasso only.