Utilities

kalelinear.utils.lap_norm(X, n_neighbour=3, metric='cosine', mode='distance', normalise=True)[source]

[summary]

Parameters:

  • X ([type]) – [description]

  • n_neighbour (int, optional) – [description], by default 3

  • metric (str, optional) – [description], by default ‘cosine’

  • mode (str, optional) – {‘connectivity’, ‘distance’}, by default ‘distance’. Type of returned matrix: ‘connectivity’ will return the connectivity matrix with ones and zeros, and ‘distance’ will return the distances between neighbors according to the given metric.

  • normalise (bool, optional) – [description], by default True

Returns:

[description]

Return type:

[type]

kalelinear.utils.mmd_coef(ns, nt, ys=None, yt=None, kind='marginal', mu=0.5)[source]
kalelinear.utils.centering_matrix(size, dtype=<class 'numpy.float64'>)[source]

Generate a centering matrix.

kalelinear.utils.centered_kernel_matrix(X, kernel='linear', metric=None, filter_params=True, **kwargs)[source]

Compute a centered kernel matrix for samples in X.

kalelinear.utils.hsic_grad_term(w, X, covariates)[source]

Compute X.T H C C.T H X w for linear-kernel HSIC regularization.

kalelinear.utils.kernel_fit_matrices(X, kernel='linear', metric=None, filter_params=True, **kwargs)[source]

Prepare common fit-time kernel, identity, and centering matrices.

kalelinear.utils.base_init(X, kernel='linear', **kwargs)[source]
kalelinear.utils.to_numpy(value)[source]