THUMedInfo/RareArena

RareArena是一个全面的罕见疾病诊断数据集，涵盖近50,000名患者和超过4,000种疾病。数据集仍在开发中，相关论文即将发布。

RareArena is a comprehensive rare disease diagnostic dataset with nearly 50,000 patients covering more than 4000 diseases. The dataset is primarily used for rare disease diagnosis, with evaluation methods including generating the top 5 diagnosis results using a model, then evaluating using GPT-4o, and finally parsing the evaluation output to calculate top-1 and top-5 recall. The data collection process requires reproducing PMC-Patients first, followed by following the pipeline described in the paper.

搜集汇总

背景与挑战

背景概述

THUMedInfo/RareArena是一个专注于罕见疾病诊断的综合数据集，包含近50,000名患者数据，覆盖超过4,000种疾病。该数据集基于PMC病例报告构建，使用GPT-4进行数据处理，适用于问答任务，采用CC BY-NC-SA 4.0许可协议。