西班牙语放射学报告标注语料库

本研究创建了一个包含513份西班牙语放射学报告的标注语料库，旨在为命名实体识别（NER）和关系抽取（RE）算法提供评估资源。数据集由阿根廷的一家医院提供，报告内容涉及多种超声检查，如肾脏和腹部检查。创建过程中，研究人员通过多次迭代改进了标注指南，并利用自动预标注技术加速了标注过程。该数据集的应用领域主要集中在医疗文本处理，特别是帮助医生识别可能的医疗问题，从而指导手术干预等治疗决策。

This study developed an annotated corpus containing 513 Spanish radiology reports, with the aim of providing evaluation resources for named entity recognition (NER) and relation extraction (RE) algorithms. The dataset was supplied by a hospital in Argentina, and the reports cover a range of ultrasound examinations, including kidney and abdominal ultrasound scans. During the construction of this corpus, researchers refined the annotation guidelines through multiple iterative rounds and employed automatic pre-annotation techniques to expedite the annotation process. The primary application scope of this dataset lies in medical text processing, specifically aiding clinicians in identifying potential medical concerns to inform therapeutic decisions such as surgical interventions.