FEEL Benchmark Code
Paper: singh-2026-feel
Code: Check arXiv paper for GitHub link (preprint: 2604.05926)
What It Provides
- Unified preprocessing pipeline for 19 EDA + PPG datasets
- 16 architecture implementations (ML, DL, pretraining, CLSP)
- Leave-one-subject-out cross-validation framework
- Cross-dataset transfer evaluation
How to Use
- Clone the repo
- Run Meltdown Minder features through their preprocessing
- Benchmark against their 16 architectures
- Use their cross-dataset evaluation for external validation
Linked To
- Dataset: feel-benchmark
- Paper: singh-2026-feel