Added an example that uses datasets from MOABB and the RG+QuantumSVM pipeline #39
Conversation
…rain Invaders datasets. It classifies using a Quantum Classifier and Riemannian geometry.
…ithub.com/toncho11/pyRiemann-qiskit into add_classification_for_BrainInvaders_2012
examples/ERP/classify_P300_bi.py
Outdated
| class Vectorizer(BaseEstimator, TransformerMixin): | ||
| def __init__(self): | ||
| pass | ||
|
|
||
| def fit(self, X, y): | ||
| """fit.""" | ||
| return self | ||
|
|
||
| def transform(self, X): | ||
| """transform. """ | ||
| return np.reshape(X, (X.shape[0], -1)) |
There was a problem hiding this comment.
If we intend to reuse this, can you add it to utils.filtering?
There was a problem hiding this comment.
Do we still need this since you changed your example?
examples/ERP/classify_P300_bi.py
Outdated
| shots=None, #'None' forces classic SVM | ||
| nfilter=2, #default 2 | ||
| dim_red=PCA(n_components=10), #default 10, higher values render better performance with the SVM version used in qiskit | ||
| #q_account_token="" #IBM Quantum TOKEN |
There was a problem hiding this comment.
Can you raise a defect for the q_account_token, following your observations in our Skype conversation?
There was a problem hiding this comment.
Follow-up opened here: #40
examples/ERP/classify_P300_bi.py
Outdated
| pipelines["Xdw+LDA"] = make_pipeline( | ||
| Xdawn(nfilter=2, estimator="scm"), | ||
| Vectorizer(), | ||
| LDA(solver="lsqr", shrinkage="auto") |
There was a problem hiding this comment.
You do not have the same number of component in your feature vector, as compared with the two other pipelines. Is this intended?
examples/ERP/classify_P300_bi.py
Outdated
| ax.set_ylabel("ROC AUC") | ||
| ax.set_ylim(0.3, 1) | ||
|
|
||
| fig.show() |
There was a problem hiding this comment.
Can you post on a comment here the resulting figure?
| from pyriemann_qiskit.classification import QuantumClassifierWithDefaultRiemannianPipeline | ||
| from sklearn.decomposition import PCA | ||
|
|
||
|
|
This comment was marked as outdated.
This comment was marked as outdated.
Sorry, something went wrong.
examples/ERP/classify_P300_bi.py
Outdated
| It uses QuantumClassifierWithDefaultRiemannianPipeline. This pipeline uses | ||
| Riemannian Geometry and Tangent Space to generate features and a quantum SVM | ||
| classifier. It uses MOABB for the evaluation and compariosn with 2 other | ||
| standard pipelines. |
There was a problem hiding this comment.
Adapt here text according to our Skype conversation.
examples/ERP/classify_P300_bi.py
Outdated
| # Here we provide some pipelines for comparison: | ||
|
|
||
| # Standard SVM on raw data | ||
| from sklearn import svm |
There was a problem hiding this comment.
Move this at the beginning.
The number of pipelines has been reduced to two. This way it will be more simple to compare results.
add `print(__doc__)` to example
pyriemann_qiskit/utils/filtering.py
Outdated
| pass | ||
|
|
||
| def fit(self, X, y): | ||
| """fit.""" |
There was a problem hiding this comment.
Can you add descriptions for Parameters and Returns, for fit() and transform() ?
pyriemann_qiskit/utils/filtering.py
Outdated
| class Vectorizer(BaseEstimator, TransformerMixin): | ||
| """This is an auxiliary transformer that allows one to vectorize data | ||
| structures in a pipeline For instance, in the case of an X with dimensions | ||
| Nt x Nc x Ns, one might be interested in a new data structure with |
There was a problem hiding this comment.
Can you use the standard names for dimensions: n_samples, n_features, n_channels, etc.
…ithub.com/toncho11/pyRiemann-qiskit into gc/add_classification_for_BrainInvaders_2012
Co-authored-by: Quentin Barthélemy <q.barthelemy@gmail.com>
…ithub.com/toncho11/pyRiemann-qiskit into add_classification_for_BrainInvaders_2012
|
Hi, I think everything is ready for this PR to be accepted. |
|
@sylvchev did you have a chance to review the last modification? |
Change the way `q_account_token` changed (updated in last PR)
|
@toncho11 thanks for taking care of this. Do you think we can reuse some elements of your presentation at |
|
I was wondering how can provides an estimate of the chance level for the bi2012 dataset. Is there something we can reuse from moabb @sylvchev ? |
The example visualizes and compares with two other "standard" pipelines.
It depends on the newest version of MOABB 0.4.6.