skada.OTLabelProp
- skada.OTLabelProp(base_estimator=None, reg=0, metric='sqeuclidean', n_iter_max=200)[source]
Label propagation using optimal transport plan.
This adapter uses the optimal transport plan to propagate labels from source to target domain. This was proposed originally in [28] for semi-supervised learning and can be used for domain adaptation.
- Parameters:
- base_estimatorobject
The base estimator to be used for the classification task. This estimator should optimize a classification loss corresponding to the given metric and provide compatible predict method (decision_function of predict_proba).
- regfloat, default=0
The entropic regularization parameter for the optimal transport problem. If None, the exact OT is solved, else it is used to weight the entropy regularizationof the coupling matrix.
- metricstr, default='sqeuclidean'
The metric to use for the cost matrix. Can be 'sqeuclidean' for squared euclidean distance, 'euclidean' for euclidean distance,
- n_iter_max: int
Maximum number of iterations for the OT solver.
- Returns:
- adapterOTLabelPropAdapter
The optimal transport label propagation adapter.
References
- [28] Solomon, J., Rustamov, R., Guibas, L., & Butscher, A. (2014, January).
Wasserstein propagation for semi-supervised learning. In International Conference on Machine Learning (pp. 306-314). PMLR.
