brainspace.gradient.gradient.GradientMaps

class brainspace.gradient.gradient.GradientMaps(n_components=10, approach='dm', kernel=None, alignment=None, random_state=None)[source]

Gradient maps.

Parameters:
  • n_components (int, optional) – Number of gradients. Default is 10.
  • approach ({'dm', 'le', 'pca'} or object, optional) –

    Embedding approach. Default is ‘dm’. It can be a string or instance:

  • kernel (str, callable or None, optional) – Kernel function to build the affinity matrix. Possible options: {‘pearson’, ‘spearman’, ‘cosine’, ‘normalized_angle’, ‘gaussian’}. If callable, must receive a 2D array and return a 2D square array. If None, use input matrix. Default is None.
  • alignment ({'procrustes', 'joint'}, object or None) –

    Alignment approach. Only used when two or more datasets are provided. If None, no alignment is performed. If object, it accepts an instance of ProcrustesAlignment. Default is None.

    • If ‘procrustes’, datasets are aligned using generalized procrustes analysis.
    • If ‘joint’, datasets are embedded simultaneously based on a joint affinity matrix built from the individual datasets. This option is only available for ‘dm’ and ‘le’ approaches.
  • random_state (int or None, optional) – Random state. Default is None.
Variables:
  • lambdas (ndarray or list of arrays, shape = (n_components,)) – Eigenvalues for each datatset.
  • gradients (ndarray or list of arrays, shape = (n_samples, n_components)) – Gradients (i.e., eigenvectors).
  • aligned (None or list of arrays, shape = (n_samples, n_components)) – Aligned gradients. None if alignment is None or only one dataset is used.
__init__(n_components=10, approach='dm', kernel=None, alignment=None, random_state=None)[source]

Initialize self. See help(type(self)) for accurate signature.

Methods

__init__([n_components, approach, kernel, …]) Initialize self.
fit(x[, gamma, sparsity, n_iter, reference]) Compute gradients and alignment.
get_params([deep]) Get parameters for this estimator.
set_params(**params) Set the parameters of this estimator.
fit(x, gamma=None, sparsity=0.9, n_iter=10, reference=None, **kwargs)[source]

Compute gradients and alignment.

Parameters:
  • x (ndarray or list of arrays, shape = (n_samples, n_feat)) – Input matrix or list of matrices.
  • gamma (float or None, optional) – Inverse kernel width. Only used if kernel == 'gaussian'. If None, gamma=1/n_feat. Default is None.
  • sparsity (float, optional) – Proportion of the smallest elements to zero-out for each row. Default is 0.9.
  • n_iter (int, optional) – Number of iterations for procrustes alignment. Default is 10.
  • reference (ndarray, shape = (n_samples, n_feat), optional) – Initial reference for procrustes alignments. Only used when alignment == 'procrustes'. Default is None.
  • kwargs (kwds, optional) – Additional keyword parameters passed to the embedding approach.
Returns:

self (object) – Returns self.