adelie.state.css_cov#

adelie.state.css_cov(S: ndarray, subset_size: int, subset: ndarray, method: str, loss: str, max_iters: int, n_threads: int)[source]#

Creates a CSS covariance state object.

Parameters:

S(p, p) ndarray

Positive semi-definite matrix \(\Sigma\).

subset_sizeint

Subset size \(k\). It must satisfy the following conditions for each method type:

"greedy": must be an integer.

"swapping": must satisfy the conditions for "greedy" if subset is None. Otherwise, it is ignored.

subsetndarray

Initial subset \(T\). This argument is only used by the swapping method. If None, the greedy method is used to first initialize a subset of size subset_size.

methodstr

Search method to identify the optimal \(T\). It must be one of the following:

"greedy": greedy method.

"swapping": swapping method.

lossstr

Loss type. It must be one of the following:

"least_squares": least squares loss.

"subset_factor": subset factor loss.

"min_det": minimum determinant loss.

max_itersint

Maximum number of cycles.

n_threadsint

Number of threads.

Returns:

wrap: Wrapper state object.