skada.TransferComponentAnalysisAdapter

skada.TransferComponentAnalysisAdapter(kernel='rbf', n_components=None, mu=0.1)

Transfer Component Analysis.

See [9] for details.

Parameters:

kernelkernel object, default='rbf'

The kernel computed between data.

n_componentsint, default=None

The numbers of components to learn with PCA. Should be less or equal to the number of samples of the source and target data.

mufloat, default=0.1

The parameter of the regularization in the optimization problem.

References

[9]

Sinno Jialin Pan et. al. Domain Adaptation via Transfer Component Analysis. In IEEE Transactions on Neural Networks, 2011.

Attributes:

`X_source_`array

Source data used for the optimization problem.

`X_target_`array

Target data used for the optimization problem.

`K_`array

Kernel distance between the data (source and target).

`eigvects_`array

Highest n_components eigenvectors of the solution of the optimization problem used to project in the new subspace.