ODD

ODD是一个用于自然语言处理基础的阿片类药物相关异常行为检测的基准数据集，由麻省大学阿默斯特分校等机构创建。该数据集包含从患者电子健康记录笔记中识别的异常行为，并将其分类为九个类别。数据集旨在通过先进的自然语言处理模型，识别与阿片类药物使用相关的风险行为，从而预防阿片类药物过量等严重后果。

ODD is a benchmark dataset for opioid-related anomalous behavior detection in foundational natural language processing (NLP) research, developed by institutions including the University of Massachusetts Amherst. This dataset contains anomalous behaviors identified from patient electronic health record (EHR) notes, which are categorized into nine classes. The dataset is designed to enable advanced natural language processing models to identify risk behaviors associated with opioid use, thereby preventing severe consequences such as opioid overdose.