This is a randomly split five fold dataset of the MIMIC-IV-ECG varaint of ECG-QA stratified based on patients (i.e., zero patient overlap between training and test).
The name of the dataset repository is ecg-qa-mimic-iv-ecg-250-1250 where 250 refers to the sampling frequency and 1250 denotes 5 seconds.
The code to do the splitting is here.
The dataset contents are under the same license as the original ECG-QA's license.
Any questions or issues, please do not hesitate… See the full description on the dataset page:
https://huggingface.co/datasets/willxxy/ecg-qa-mimic-iv-ecg-250-1250.
