Add Riemann Quantum Classifier #31
Conversation
add some tests to riemann quantum classifier
pyriemann_qiskit/classification.py
Outdated
| nfilter : int (default: 1) | ||
| The number of filter for the xDawnFilter. | ||
| The number of components selected is 2 x nfilter. | ||
| dim_red : TransformerMixin (default: NaiveDimRed(is_even=True)) |
There was a problem hiding this comment.
PCA seems a better default choice, because NaiveDimRed can easily loose discriminant features.
pyriemann_qiskit/classification.py
Outdated
| @@ -417,3 +422,192 @@ def predict(self, X): | |||
| """ | |||
| _, labels = self._predict(X) | |||
| return self._map_0_1_to_classes(labels) | |||
|
|
|||
|
|
|||
| class RiemannQuantumClassifier(BaseEstimator, ClassifierMixin, | |||
There was a problem hiding this comment.
Not sure that the current name is adapted.
The method looks like a full pipeline with strong constraints (Xdawn, tangent space, dimension reduction), rather than a new generic classifier.
There was a problem hiding this comment.
I agree. I will try to find something more appropriate.
pyriemann_qiskit/datasets/utils.py
Outdated
| """Return mne epochs with duration 1s. | ||
| Only visual left and right trials are selected. | ||
| Data are returned filtered. |
There was a problem hiding this comment.
Can you complete the description of the task of the signal, and add the link towards MNE dataset page?
Co-authored-by: Quentin Barthélemy <q.barthelemy@gmail.com>
|
Thanks for the review! There is still a problem with the pipeline, using python 3.8. To be honest I am not sure where it comes from. |
|
Ok, I compared the package version with older builds and it seems that the scipy version changed 2 days ago. @toncho11 FYI |
|
Thx @gcattan ! |
|
Thanks for the review :) |
This PR implements #23.
The helper classifier which is implemented here will be used for CVGridSearch.
A couple of tests were added to
test_classification.pyand required to refactorize some of the methods into a dataset/utils module.(WIP: missing documentation of the dataset module, and error in the pipeline)