moabb.evaluations.CrossSubjectSplitter#

class moabb.evaluations.CrossSubjectSplitter(cv_class: type[~sklearn.model_selection._split.BaseCrossValidator] = <class 'sklearn.model_selection._split.LeaveOneGroupOut'>, random_state: int | None = None, **cv_kwargs)[source]#

Data splitter for cross subject evaluation.

This splitter enables cross-subject evaluation by performing a Leave-One-Session-Out (LOSO) cross-validation on the dataset.

It assumes that the entire metainformation across all subjects is already loaded.

Unlike the CrossSubjectEvaluation class from moabb.evaluation, which manages the complete evaluation process end-to-end, this splitter is solely responsible for dividing the data into training and testing sets based on subjects.

The schematic diagram of the CrossSubject split

The splitting strategy for the subjects can be changed by passing the cv_class and cv_kwargs arguments. By default, it uses LeaveOneGroupOut, which performs Leave-One-Subject-Out cross-validation.

Parameters:

  • cv_class (cross-validation class, default=LeaveOneGroupOut) – Cross-validation strategy for splitting the subjects between train and test sets. By default, use LeaveOneGroupOut, which keeps one subject as a test.

  • random_state (int, RandomState instance or None, default=None) – Controls the randomness of the cross-validation. Pass an int for reproducible output across multiple calls.

  • cv_kwargs (dict) – Additional arguments to pass to the inner cross-validation strategy.

Yields:

  • train (ndarray) – The training set indices for that split.

  • test (ndarray) – The testing set indices for that split.

get_n_splits(metadata)[source]#

Return the number of splits for the cross-validation.

The number of splits is the number of subjects times the number of splits of the inner cross-validation strategy.

We try to keep the same behaviour as the sklearn cross-validation classes.

Parameters:

metadata (pd.DataFrame) – The metadata containing the subject and session information.

Returns:

n_splits – The number of splits for the cross-validation

Return type:

int

split(y, metadata)[source]#

Generate indices to split data into training and test set.

Parameters:

  • X (array-like of shape (n_samples, n_features)) – Training data, where n_samples is the number of samples and n_features is the number of features.

  • y (array-like of shape (n_samples,)) – The target variable for supervised learning problems.

  • groups (array-like of shape (n_samples,), default=None) – Group labels for the samples used while splitting the dataset into train/test set.

Yields:

  • train (ndarray) – The training set indices for that split.

  • test (ndarray) – The testing set indices for that split.