Persius/imr-appeals

该数据仓库包含了手动标记和伪标记的背景跨度，这些数据对应于美国健康保险案例中的外部上诉裁决，用于覆盖拒绝的情况。数据集的具体描述和更详细的工作文档可以在提供的GitHub链接中找到。

This dataset contains manually labeled and pseudo-labeled background spans corresponding to real external appeal adjudications, specifically coverage denials, from U.S. health insurance cases. The dataset features include text, decision, appeal type, full text, and sufficiency ID. The dataset is divided into training and test sets, containing 64067 and 9920 samples respectively. The download size of the dataset is 92577490 bytes, and the dataset size is 199523278 bytes.