MIMIC-IV-ECG

该数据集是一个子集，旨在预测病人在接下来的24小时内是否需要重症监护室（ICU）的收治。此外，基于差分隐私（DP）和真正阳性率（TPR）指标，已经进行了公平性分析，该分析揭示了不同人口统计学群体之间的差异。该任务的目标是预测24小时内是否需要ICU入院。

This dataset is a subset aimed at predicting whether a patient will require intensive care unit (ICU) admission within the next 24 hours. Additionally, a fairness analysis has been conducted based on Differential Privacy (DP) and True Positive Rate (TPR) metrics, which has revealed disparities across different demographic groups. The goal of this task is to predict the need for ICU admission within 24 hours.

搜集汇总

背景与挑战

背景概述

MIMIC-IV-ECG是一个大规模诊断性心电图数据集，包含约80万份12导联心电图，覆盖16万患者，与MIMIC-IV临床数据库匹配，支持跨模块数据关联研究。数据集提供原始波形、机器测量指标及部分心电图对应的 cardiologist 报告链接，采样率为500Hz，时间跨度为2008-2019年。