肝癌领域精准医学语料标注研究

摘要

目的 构建高质量的肝癌相关精准医学语料,缓解当前生物医学语料库紧缺的问题.方法 根据已有生物医学语料库构建方法制定初步标注规范,利用Brat工具预标注迭代完善规范,最终在该规范指导下进行命名实体和实体关系的正式标注.结果 完成了210篇摘要的语料标注,标注了命名实体10045个以及语义关系2489个.结论 形成质量较高的肝癌领域精准医学标注语料,制定了精准医学语料标注规范.%Objective To construct a high-quality precision medicine corpus in liver neoplams and to alleviate the shortage of biomedical corpora .Methods Initial corpus annotation guideline were drafted according to current biomedical corpora, and the guidelines were gradually revised through pre-annotation using Brat tool before formal annotation of the named entities and the relations was started .Results The corpus annotation of 210 abstracts was completed , including 10045 named entities and 2489 semantic relations .Conclusion A high-quality precision medicine corpus in liver neoplasmis made available and the annotation guidelines are formulated .