Quentin Barthélemy
Quentin Barthélemy
This PR enhances AJDC when used for group BBS (gBSS): it allows to refine group sources for a specific subject, using an initial diagonalizer. - Add parameter 'init' to `ajd_pham`...
In P300 datasets, each trial is usually associated to a (binary) _ERP class_ : target (target character has been flashed, evoking an ERP in the trial) versus non-target (target character...
This PR adds parameters to `random_point()`, in order to have more flexibility when generating random SPD matrices. Default values give the same behavior as before.
#### What does this implement/fix? Real-time EEG based systems like BCI or neurofeedback use only causal filters, specially forward IIR filters to avoid delay. Replaying data in offline mode, we...
Many classes and functions check `metric` argument, but each module has its own function (classification, tangentspace, transfer). This PR: - uses a single function, - completes docs giving the link...
## Description A [Hermitian positive definite (HPD) matrix](https://en.wikipedia.org/wiki/Definite_matrix#Definitions_for_complex_matrices) $M$ is defined as: $M = A + iB$ where $A$ is a symmetric positive definite (SPD) matrix and $B$ is a...
This PR aims to: - add a generic `ajd` function, solving partially #49; - enhance functions `rjd` and `uwedge`, removing np.dot and np.diag(np.diag()), and using uniform notations; - enhance `ajd_pham`,...
When the tests of the different variables have not the same degrees of freedom, it is not possible to estimate the joint null distribution using the max (or min) of...
This PR: improves code of tests, transforms classes into functions into clustering and embedding tests.
This PR enhances `ajd_pham` to process HPD matrices, following code https://github.com/Marco-Congedo/Diagonalizations.jl/blob/master/src/optim/LogLike.jl#L46 Note: `tau` in pyRiemann is the transpose of `Gamma` in Diagonalizations.jl This is the continuation of #238. Related to...