mcmodels.models.homogeneous.backward_subset_selection_conditioning

mcmodels.models.homogeneous.backward_subset_selection_conditioning(X, kappa=1000)

Conditioning through subselecting columns of X.

Uses svd_subset_selection to greedily remove single columns of X till the desired conditioning is reached.

Parameters:

X : array, shape (n_samples, n_features)

Array whose columns we wish to subset.

kappa : float, optional, default: 1000

The maximum condition number desired.

Returns:

C : array, shape (n_samples, ?)

Array with condition number < kappa, containing the lesser linearly dependent columns of X.