pyntbci.stopping.BayesStopping

class pyntbci.stopping.BayesStopping(estimator: ClassifierMixin, segment_time: float, fs: int, method: str = 'bds0', cr: float = 1.0, target_pf: float = 0.05, target_pd: float = 0.8, max_time: float = None, min_time: float = None, approach: str = 'template_inner')

Bayesian dynamic stopping. Fits Gaussian distributions for target and non-target responses, and calculates a stopping threshold using these and a cost criterion [1].

Parameters:

• estimator (ClassifierMixin) – The classifier object that performs the classification.

• segment_time (float) – The size of a segment of data at which classification is performed ins seconds.

• fs (int) – The sampling frequency of the EEG data in Hz.

• method (str (default: "bds0")) – The method to use for Bayesian dynamic stopping: bds0, bds1, bds2.

• cr (float (default: 1.0)) – The cost ratio.

• target_pf (float (default: 0.05)) – The targeted probability for error.

• target_pd (float (default: 0.80)) – The targeted probability for detection.

• max_time (float (default: None)) – The maximum time in seconds at which to force a stop, i.e., a classification. Trials will not be longer than this maximum time. If None, the algorithm will always emit -1 if it cannot stop.

• min_time (float (default: None)) – The minimum time in seconds at which a stop is possible, i.e., a classification. Before the minimum time, the algorithm will always emit -1. If None, the algorithm allows a stop already after the first segment of data.

• approach (str (default: "template_inner")) – The approach used to fit the BDS model. Either an analytic template-based method using the inner product, or an approach using the empirical scores from the estimator object.

alpha_

The scaling parameter between observed and predicted responses.

Type:

float

sigma_

The standard deviation of the noise.

Type:

float

b0_

The mean of the non-target Gaussian distribution of shape (n_segments).

Type:

NDArray

b1_

The mean of the target Gaussian distribution of shape (n_segments).

Type:

NDArray

s0_

The standard deviation of the non-target Gaussian distribution of shape (n_segments).

Type:

NDArray

s1_

The standard deviation of the target Gaussian distribution of shape (n_segments).

Type:

NDArray

eta_

The decision boundary of shape (n_segments).

Type:

NDArray

pf_

The predicted probability of a false detection of shape (n_segments).

Type:

NDArray

pm_

The predicted probability of a miss of shape (n_segments).

Type:

NDArray

References

[1]

Ahmadi, S., Desain, P., & Thielen, J. (2024). A Bayesian dynamic stopping method for evoked response brain-computer interfacing. Frontiers in Human Neuroscience, 18, 1437965.

fit(X: ndarray[tuple[Any, ...], dtype[_ScalarT]], y: ndarray[tuple[Any, ...], dtype[_ScalarT]]) → ClassifierMixin

The training procedure to fit the dynamic procedure on supervised EEG data.

Parameters:

• X (NDArray) – The matrix of EEG data of shape (n_trials, n_channels, n_samples).

• y (NDArray) – The vector of ground-truth labels of the trials in X of shape (n_trials).

Returns:

self – Returns the instance itself.

Return type:

ClassifierMixin

predict(X: ndarray[tuple[Any, ...], dtype[_ScalarT]]) → ndarray[tuple[Any, ...], dtype[_ScalarT]]

The testing procedure to apply the estimator to novel EEG data using Bayesian dynamic stopping.

Parameters:

X (NDArray) – The matrix of EEG data of shape (n_trials, n_channels, n_samples).

Returns:

y – The vector of predicted labels of the trials in X of shape (n_trials). Note, the value equals -1 if the trial cannot yet be stopped.

Return type:

NDArray

set_score_request(*, sample_weight: bool | None | str = '$UNCHANGED$') → BayesStopping

Configure whether metadata should be requested to be passed to the score method.

Note that this method is only relevant when this estimator is used as a sub-estimator within a meta-estimator and metadata routing is enabled with enable_metadata_routing=True (see sklearn.set_config()). Please check the User Guide on how the routing mechanism works.

The options for each parameter are:

• True: metadata is requested, and passed to score if provided. The request is ignored if metadata is not provided.

• False: metadata is not requested and the meta-estimator will not pass it to score.

• None: metadata is not requested, and the meta-estimator will raise an error if the user provides it.

• str: metadata should be passed to the meta-estimator with this given alias instead of the original name.

The default (sklearn.utils.metadata_routing.UNCHANGED) retains the existing request. This allows you to change the request for some parameters and not others.

Added in version 1.3.

Parameters:

sample_weight (str, True, False, or None, default=sklearn.utils.metadata_routing.UNCHANGED) – Metadata routing for sample_weight parameter in score.

Returns:

self – The updated object.

Return type:

object