VitalDBPreprocessed / README.md
nbwolfer's picture
Add dataset card
a62cc9d verified
|
Raw History Blame Contribute Delete
3.15 kB
---
license: cc-by-4.0
pretty_name: VitalDB Preprocessed
language:
- en
tags:
- medical
- biosignals
- vital-signs
- time-series
- perioperative
- multimodal
task_categories:
- time-series-forecasting
---
# VitalDB Preprocessed
A preprocessed version of **VitalDB**, an open database of intraoperative vital signs. It is prepared for research on temporal, multimodal patient-state modelling (digital twins) from asynchronous clinical observations.
> **Status:** no data has been published yet. This card will be updated with the preprocessing recipe, schema and revision when the first prepared version is uploaded.
## Source
- **Original dataset:** VitalDB, a high-fidelity multi-parameter vital signs database in surgical patients. 6,388 surgical cases from Seoul National University Hospital, with waveform and numeric monitoring tracks, perioperative clinical parameters and time-series laboratory results.
- **Distribution used:** [PhysioNet, version 1.0.0](https://physionet.org/content/vitaldb/1.0.0/)
- **Original license:** [Creative Commons Attribution 4.0 International](https://creativecommons.org/licenses/by/4.0/)
This derived dataset is redistributed under the same CC BY 4.0 license. It is not endorsed by the VitalDB authors, Seoul National University Hospital or PhysioNet.
## Changes from the original
To be filled in with the first published version:
- Cases and tracks selected
- Resampling and alignment of waveform, numeric and laboratory streams
- Handling of missing values and gaps
- Split method (by case, so no case appears in more than one split)
- Output format and shard layout
The original case IDs are kept so every record can be traced back to VitalDB. No information beyond the original de-identified release has been added. Do not attempt to re-identify individuals.
## Usage
Each prepared version is a Git commit of this repository with a `manifest.json` listing every shard with its size and SHA-256. Pin experiments to a full commit SHA, not to `main`.
```python
from huggingface_hub import snapshot_download
path = snapshot_download(
"nbwolfer/VitalDBPreprocessed",
repo_type="dataset",
revision="<full 40-character commit SHA>",
)
```
## Citation
If you use this dataset, cite the original VitalDB publication and the PhysioNet release:
```bibtex
@article{lee2022vitaldb,
title = {VitalDB, a high-fidelity multi-parameter vital signs database in surgical patients},
author = {Lee, Hyung-Chul and Park, Yoonsang and Yoon, Soo Bin and Yang, Seong Mi and Park, Dongnyeok and Jung, Chul-Woo},
journal = {Scientific Data},
volume = {9},
pages = {279},
year = {2022},
doi = {10.1038/s41597-022-01411-5}
}
@misc{vitaldb_physionet,
title = {VitalDB, a high-fidelity multi-parameter vital signs database in surgical patients (version 1.0.0)},
publisher = {PhysioNet},
year = {2022},
doi = {10.13026/czw8-9p62}
}
```
PhysioNet also asks users to cite: Goldberger, A., et al. (2000). PhysioBank, PhysioToolkit, and PhysioNet: Components of a new research resource for complex physiologic signals. *Circulation*, 101(23), e215–e220.