10,000 RR Interval Data (9500NAF & 500PAF) from 24 h Holter recordings used for atrial fibrillation detection

This RR interval dataset is derived from 10,000 cases of 24-hour Holter monitoring data sampled at 128 Hz. Among the cases, 9,500 are labeled as non-atrial fibrillation (NAF), and 500 as paroxysmal atrial fibrillation (PAF). These data have been used in the article "Clinician-AI Collaboration: A Win-Win solution for Efficiency and Reliability in Atrial Fibrillation Diagnosis" published in Med. https://doi.org/10.1016/j.medj.2025.100668The dataset formated as CSV file consists of two columns:<b>rr_interval</b>: Represents the interval between consecutive R-peaks, measured in milliseconds.<b>label</b>: Categorical labels for the beats, where:<code>1</code> indicates AF<code>0</code> indicates NAF<code>-1</code> indicates noise or artifactsEach case is named based on its category. NAF cases are labeled as <code>NAF0001.csv</code> through <code>NAF9500.csv</code>, while PAF cases are labeled as <code>PAF0001.csv</code> through <code>PAF0500.csv</code>.If you use this dataset, please cite the following article:Zhang P, Lin F, Ma F, Chen Y, Liu Y, Feng X, Fang S, Zhang H, Xiao S, Yang X, Li D, Wang DW, Yang X, Li Q. Clinician-artificial intelligence collaboration: A win-win solution for efficiency and reliability in atrial fibrillation diagnosis. Med. 2025 Jul 11;6(7):100668. doi: 10.1016/j.medj.2025.100668.For any questions, please contact the email: [e-mail]