| --- |
| license: mit |
| language: |
| - en |
| tags: |
| - ecg |
| - arrhythmia |
| - healthcare |
| - time-series |
| - biosignal |
| --- |
| |
| # ECG Arrhythmia Dataset — MIT-BIH Processed |
|
|
| Preprocessed beat segments derived from the MIT-BIH Arrhythmia Database (PhysioNet). |
|
|
| ## Contents |
|
|
| | File | Description | |
| |---|---| |
| | X.npy | Beat segments — shape (N, 180), float32 | |
| | y.npy | Encoded labels — integers 0-4 | |
| | y_raw.npy | Raw string labels — N, L, R, V, A | |
| |
| ## Classes |
| |
| | Label | Class | Description | |
| |---|---|---| |
| | 0 | N | Normal sinus rhythm | |
| | 1 | L | Left bundle branch block | |
| | 2 | R | Right bundle branch block | |
| | 3 | V | Premature ventricular contraction | |
| | 4 | A | Atrial premature contraction | |
| |
| ## Source |
| Raw data from [MIT-BIH Arrhythmia Database](https://physionet.org/content/mitdb/1.0.0/). |
| Preprocessing: bandpass filter (0.5–50 Hz) + z-score normalization + 180-sample windowing. |
| |
| ## Related |
| - [Model](https://huggingface.co/dheerajthuvara/ecg-arrhythmia-detection) |
| - [GitHub](https://github.com/dheerajthuvara/ecg-arrhythmia-detection) |
| - [Portfolio](https://dtlabs.me) |
