External Email - Use Caution
Hi everyone,
I've done CSP(Common Spatial Patterns) and classification for epoched
EEG-Data. I'm thinking about segmenting (cutting) each epoch into sgments,
where each segment consists of about 10 Frames. In this way I can increase
the data. So let's say I have 5*epochs from each class. And let's say each
epoch has 200 Frames. I cut each epoch into 20 Segments. Each segment has
10 Frames. In this case, I have instead of 5 trials from each class, I have
now 5*20 trials. Is it possible in the CSP? Does it make sense to do so?
I'm asking because I saw they use this method in Deep Learning, where they
cut the data into segments and then feed them into convolutional +
recurrent neural networks.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://mail.nmr.mgh.harvard.edu/pipermail/mne_analysis/attachments/20190307/80aa9533/attachment.html