How sharan-yan04/eeg_classifier works
This repository presents a notebook-driven EEG motor-imagery classification workflow: obtain PhysioNetMI trials, extract filter-bank CSP features, select features with QGA, and classify with an SVM. The README also describes evaluation statistics and a Gradio demo supporting staged examples and manual EEG input. The tree identifies one main notebook, but its contents were not sampled; component paths and relationships below therefore follow the README’s documented architecture, not inspected notebook calls.
Main entry points: QGA_for_FBCSP_feature_selection_(Final).ipynb Gradio demo
Components 13
EEG data
- PhysioNetMI
- EEG trials
notebooks/QGA_for_FBCSP_feature_selection_(Final).ipynb
Feature pipeline
- Three-band filterbank
notebooks/QGA_for_FBCSP_feature_selection_(Final).ipynb - FBCSP features
notebooks/QGA_for_FBCSP_feature_selection_(Final).ipynb - Train-only scaling
notebooks/QGA_for_FBCSP_feature_selection_(Final).ipynb
Selection and classification
- QGA selection
notebooks/QGA_for_FBCSP_feature_selection_(Final).ipynb - RBF SVM
notebooks/QGA_for_FBCSP_feature_selection_(Final).ipynb
Evaluation and demo
- Class predictions
notebooks/QGA_for_FBCSP_feature_selection_(Final).ipynb - Statistical evaluation
notebooks/QGA_for_FBCSP_feature_selection_(Final).ipynb - Gradio demo
notebooks/QGA_for_FBCSP_feature_selection_(Final).ipynb - Staged samples
notebooks/QGA_for_FBCSP_feature_selection_(Final).ipynb - Manual EEG input
notebooks/QGA_for_FBCSP_feature_selection_(Final).ipynb
Other components
- User
Connections 12
- PhysioNetMI → EEG trials: provides trials
README.md - EEG trials → Three-band filterbank: filters into bands
README.md - Three-band filterbank → FBCSP features: extracts features
README.md - FBCSP features → Train-only scaling: scales features
README.md - Train-only scaling → QGA selection: selects subset
README.md - QGA selection → RBF SVM: feeds classifier
README.md - RBF SVM → Class predictions: predicts classes
README.md - Class predictions → Statistical evaluation: evaluates results
README.md - Class predictions → Gradio demo: presents predictions
README.md - User → Gradio demo: uses interface
README.md - Staged samples → Gradio demo: supplies examples
README.md - Manual EEG input → Gradio demo: accepts trial
README.md