# Evidence and source lock

## Benchmark format

- EdgeBench / SForge repository: https://github.com/ByteDance-Seed/EdgeBench
- Documentation snapshot commit: recorded in `provenance/source_lock.json`.
- Relevant contract: `docs/en/tasks/integration-guide.md` and `docs/en/tasks/parsers.md`.

## Scientific paper

Philip de Chazal, Maria O'Dwyer, and Richard B. Reilly. “Automatic Classification of Heartbeats Using ECG Morphology and Heartbeat Interval Features.” *IEEE Transactions on Biomedical Engineering* 51(7), 1196–1206 (2004). DOI: https://doi.org/10.1109/TBME.2004.827359

The task uses the paper's patient-independent DS1/DS2 protocol and AAMI-style superclass mapping. It does not claim to reproduce the paper's proprietary implementation or exact reported performance.

## Dataset

- MIT-BIH Arrhythmia Database, version 1.0.0: https://physionet.org/content/mitdb/1.0.0/
- Version DOI: https://doi.org/10.13026/C2F305
- Dataset paper: Moody GB, Mark RG. “The impact of the MIT-BIH Arrhythmia Database.” *IEEE Engineering in Medicine and Biology Magazine* 20(3):45–50 (2001). PMID 11446209.
- License: Open Data Commons Attribution License v1.0.

PhysioNet documents 48 half-hour, two-channel ambulatory ECG recordings from 47 subjects, sampled at 360 Hz, with roughly 110,000 beat annotations adjudicated by cardiologists. The task excludes the four paced records used neither in DS1 nor DS2 and downloads files only from the versioned `1.0.0` endpoint.

## Boundary

This package is a locally constructed, evidence-backed EdgeBench-compatible candidate. It has not been reviewed, accepted, or scored by the EdgeBench authors. It is not a medical device and its outputs must not be used for diagnosis or patient care.