[Eeglablist] Multi class CSP and feature extraction

Delorme, Arnaud adelorme at ucsd.edu
Sun Oct 13 10:16:05 PDT 2019


Dear Abdul,

Not an expert in CSP but I believe the standard approach is to do pairwise CSP between your categories then use the CSP features as input to a multi-class classifier - such as softmax logistic regression. I think BCILAB can do that https://sccn.ucsd.edu/wiki/BCILAB

Found these papers if that helps

https://arxiv.org/abs/1802.09046

https://ieeexplore.ieee.org/document/4473042

Best wishes, 

Arno

>> On Oct 1, 2019, at 12:57 PM, Md. Abdul Awal <awalece04 at yahoo.com> wrote:
>> 
>> Dear EEGLablsit
>> I am trying to apply Common Spatial Pattern (CSP) in EEG Seizure prediction. The true classes are seizure, burst, CNV etc. So this is a multi-class problem. 
>> I know there are several implementation of Binary CSP available. But  I need Multi class implementation of CSP or its updated variants with feature extraction from multiclass CSP. 
>> Will be grateful if anyone can send me such implementation. Many thanks in advance. 
>> 
>> with regards, 
>> ------------------------ 
>> Dr. Md. Abdul Awal 
>> Faculty memberElectronics and Communication Engineering DisciplineKhulna University, Khulna-9208, Bangladesh Cell # +880 1786 196913
>> _______________________________________________
>> Eeglablist page: http://sccn.ucsd.edu/eeglab/eeglabmail.html
>> To unsubscribe, send an empty email to eeglablist-unsubscribe at sccn.ucsd.edu
>> For digest mode, send an email with the subject "set digest mime" to eeglablist-request at sccn.ucsd.edu
> 




More information about the eeglablist mailing list