Skip to content

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

  1. Clone the repo
  2. Run Meltdown Minder features through their preprocessing
  3. Benchmark against their 16 architectures
  4. Use their cross-dataset evaluation for external validation

Linked To