遇见数据集

CT-EBM-SP - Corpus of Clinical Trials for Evidence-Based-Medicine in Spanish (version 3)

收藏
Zenodo2026-02-02 更新2026-05-26 收录
官方服务:

资源简介:

A collection of 1200 texts (292173 tokens) about clinical trials studies and clinical trials announcements in Spanish: - 500 abstracts from journals published under a Creative Commons license, e.g. available in PubMed or the Scientific Electronic Library Online (SciELO).- 700 clinical trials announcements published in the European Clinical Trials Register and Repositorio Español de Estudios Clínicos. Texts were annotated with the following entities types: - Semantic groups from the Unified Medical Language System: ANAT, CHEM, DEVI, DISO, LIVB, PHYS and PROC.- Medical drug information: Contraindicated, Dose_or_Strength, Form, and Route_or_Mode_of_administration.- Temporal expressions: Age, Date, Duration_or_Interval, Frequency and Time.- Miscellaneous medical entities: Concept, Food_or_Drink, Observation_or_Finding, Quantifier_or_Qualifier, and Result_or_Value.- Negation/Speculation: Neg_cue, Negated, Spec_cue and Speculated.- Attributes of temporality (Future, Family_history_of, and History_of), experiencer (Patient, Family_member and Other) and other information (Hypothetical). In addition, the following semantic relationships were annotated: - Intervention-related relations: • Has_Dose_or_Strength • Has_Drug_Form • Has_Route_or_Mode • Combined_with • Used_for • Has_Result_or_Value- Temporal relations: • Before • After • Overlap • Has_Age • Has_Frequency • Has_Duration_or_Interval- Event-related relations: • Causes • Experiences • Has_Quantifier_or_Qualifier • Location_of- Assertion relations: • Negation • Speculation 81.75% of the total entities were normalized to Unified Medical Language System (UMLS) Concept Unique Identifiers (CUIs). This is the final version with the corrections made after each file was reviewed by a a second reviewer. Two annotators reviewed each corpus file. Relation extraction Python code is available at the companion GitHub repository: https://github.com/lcampillos/ct-ebm-sp-v3

本数据集包含1200篇西班牙语文本(共计292173个Token),内容围绕临床试验研究与临床试验公告展开: - 500篇来自采用知识共享(Creative Commons)许可的期刊摘要,可于PubMed或西班牙科学电子图书馆(SciELO)获取。 - 700篇临床试验公告,发布于欧洲临床试验注册库(European Clinical Trials Register)与西班牙临床试验资料库(Repositorio Español de Estudios Clínicos)。 文本已针对以下实体类型完成标注: 1. 统一医学语言系统(Unified Medical Language System, UMLS)语义组:解剖结构(ANAT)、化学物质(CHEM)、异常表现(DEVI)、疾病(DISO)、活体生物(LIVB)、物理实体(PHYS)与操作过程(PROC)。 2. 医疗药物信息:禁忌证(Contraindicated)、剂量或浓度(Dose_or_Strength)、剂型(Form)、给药途径(Route_or_Mode_of_administration)。 3. 时间表达式:年龄(Age)、日期(Date)、持续时间或间隔(Duration_or_Interval)、频率(Frequency)与时刻(Time)。 4. 其他医疗实体:医学概念(Concept)、食物或饮品(Food_or_Drink)、观察或发现(Observation_or_Finding)、限定词或修饰语(Quantifier_or_Qualifier)与结果或数值(Result_or_Value)。 5. 否定与推测:否定提示词(Neg_cue)、被否定实体(Negated)、推测提示词(Spec_cue)与被推测实体(Speculated)。 6. 时态属性(未来、家族病史、既往病史)、体验者(患者、家族成员、其他)及其他信息(假设性)。 此外,数据集还标注了以下语义关系: - 干预相关关系: • Has_Dose_or_Strength(具有剂量或浓度) • Has_Drug_Form(具有剂型) • Has_Route_or_Mode(具有给药途径) • Combined_with(联合使用) • Used_for(用于治疗) • Has_Result_or_Value(具有结果或数值) - 时间相关关系: • Before(早于) • After(晚于) • Overlap(重叠) • Has_Age(对应年龄) • Has_Frequency(对应频率) • Has_Duration_or_Interval(对应持续时间或间隔) - 事件相关关系: • Causes(导致) • Experiences(经历) • Has_Quantifier_or_Qualifier(具有限定词或修饰语) • Location_of(所在位置) - 断言关系: • Negation(否定) • Speculation(推测) 总实体中81.75%已归一化至统一医学语言系统(UMLS)概念唯一标识符(Concept Unique Identifiers, CUIs)。 本版本为最终定稿,已完成每篇文件经第二位审核人员复核后的修正工作。 每份语料库文件均经过两名标注人员审核。 关系抽取的Python代码可从配套GitHub仓库获取:https://github.com/lcampillos/ct-ebm-sp-v3

提供机构:
Zenodo
创建时间:
2025-12-29
二维码
社区交流群
二维码
科研交流群
商业服务