Dear people,
I am a newbie in EEG signal processing and classification. I have recorded EEG signal for some subjects. By far, I have done pre-processing and artifact elimination, feature extraction, and classification for just one subject. My question is about different subjects. I am wondering if I should repeat the mentioned steps for each subject individually and finally average accuracy or should I join all features obtained from different subjects and use them for classification. May I have your help?
Best Regards,
Bahareh Bamdad
I did not work in EEG for a long time, but I will give you my advice. You should use each subject alone because the EEG data is highly subject-dependent. I tried to use leave-one-out classifiers to do cross-subject classifications (train a classifier using all subjects except one and test the left subject using the classifier), but the results were inferior to each subject's model.
Check this paper
Conference Paper P300-based applications for interacting with smart mobile devices
- Badges
- Science topic
- Similar topics
- Data Mining
- Classification