EEG Sleep Stage Classification
EEG-based sleep stage classification using Transformers, cGAN, and DeltaEncoder.
Developed deep learning models for classifying EEG-based sleep stages using advanced generative techniques.
- Analyzed EEG sleep data from the International Institute for Integrative Sleep Medicine (IIIS).
- Applied generative models (cGAN, DeltaEncoder) to address class imbalance in sleep datasets.
- Built a Transformer-based model achieving 75% F1 score in multi-class sleep stage classification.
- Augmented minority classes with synthetic data, leading to a performance boost of 5–20%.
- Optimized hyperparameters automatically using Optuna for improved model efficiency.