疫情文本标注数据集

为支撑重大疫情态势分析与智能感知模型的训练与评测，本数据集采集多权威渠道的疫情相关新闻文本，经标准化意图标注处理及标准化实体关系标注处理构建而成。数据时间范围覆盖2020-2024年，无明确空间地域限制，覆盖全球范围内重大疫情相关的公开网络信息。数据采集源包括主流新闻媒体、政府疫情通报网站及公共卫生机构官网等权威渠道，依托高性能计算服务器与网络爬虫系统，采用定向爬虫抓取+自动模型初标+人工复核的方法，通过关键词过滤、主题聚类及数据清洗等质控措施保障数据质量。数据集主要包含新闻或公告原文、发布标题及实体关系标签与疫情意图标签等数据，格式规范、标注精准，可直接用于自然语言处理模型训练、知识图谱构建、疫情态势分析等研究，具有重要的科研与应用价值。This dataset is developed to support the training and evaluation of models for major epidemic situation analysis and intelligent perception. It is constructed by collecting epidemic-related news texts from multiple authoritative channels, followed by standardized intent annotation and standardized entity-relation annotation processing. The data covers the time frame from 2020 to 2024, with no explicit spatial or regional constraints, and encompasses publicly available online information related to major epidemics worldwide. The data collection sources include authoritative channels such as mainstream news media, government epidemic notification websites, and official websites of public health institutions. Leveraging high-performance computing servers and web crawler systems, the dataset adopts a three-stage workflow: targeted web crawling, automatic initial annotation via models, and manual review. Data quality is ensured through quality control measures including keyword filtering, topic clustering, and data cleaning. The dataset primarily includes original news or announcements, release titles, entity-relation labels, epidemic intent labels, and other related data. With standardized formats and accurate annotations, it can be directly utilized for research tasks such as natural language processing model training, knowledge graph construction, and epidemic situation analysis, holding significant scientific research and application value.

搜集汇总

背景与挑战

背景概述

疫情文本标注数据集是一个面向重大疫情态势分析与智能感知模型训练的专业数据集，覆盖2020-2024年全球范围的疫情相关新闻文本，数据来源于主流媒体、政府通报和公共卫生机构等权威渠道，并经过标准化意图和实体关系标注处理。该数据集采用定向爬虫抓取结合自动初标与人工复核的方法确保质量，包含新闻原文、标题及标签，格式规范，可直接用于自然语言处理模型训练、知识图谱构建和疫情分析等科研应用。