search

CHEN-XDU / GMSS

This is an implentation of paper GMSS: Graph-Based Multi-Task Self-Supervised Learning for EEG Emotion Recognition
54 stars 10 forks source link

question regarding the experimental Protocol #4

Open biannachen opened 1 year ago

biannachen commented 1 year ago

I’m trying to reproduce the results of the article using the SEED-IV dataset. I have a question about the experimental protocol for Subject-independent experiment on SEED-IV dataset (15 subjets×3 sessions×24 trials). I would like to know the leave-one-subject-out cross-validation strategy used in the article is option A or option B or others.

Option A: (Each experiment uses a section as a test set, with a total of 15*3 experiments)

Option B: (Each experiment uses three section as a test set, with a total of 15 experiments)

L-guangQ commented 1 year ago

I have the same problem, did you reproduce it? How did you do it?