病毒-宿主交互数据集

病毒-宿主交互数据集，由亚利桑那州立大学创建，包含1104篇关于病毒检测结果的科学出版物摘要。数据集通过系统文献综述收集，涵盖1940至2015年间发表的文章，涉及586种已知有哺乳动物宿主的病毒。数据集创建过程中，排除了实验感染、动物园及特定家养哺乳动物相关的文章。该数据集主要用于训练未来NER模型，以自动化提取科学出版物中的病毒-宿主检测方法，进而预测病毒溢出风险，对理解新发传染病具有重要意义。

Virus-host interaction dataset, created by Arizona State University, contains 1,104 scientific publication abstracts detailing virus detection outcomes. Collected via systematic literature review, the dataset covers articles published between 1940 and 2015, and encompasses 586 viruses known to have mammalian hosts. During the dataset curation process, articles related to experimental infections, zoo animals, and specific domestic mammals were excluded. This dataset is primarily designed for training future NER models to automatically extract virus-host detection methods from scientific publications, thus enabling the prediction of virus spillover risks, which holds critical significance for understanding emerging infectious diseases.