基于图神经网络的药物相互作用预测方法研究综述  

A review of prediction methods of drug interaction based on graph neural network

作　　者：郭文博 龙伟 蒋林华 叶文标 胡灵犀 彭司华 Guo Wenbo;Long Wei;Jiang Linhua;Ye Wenbiao;Hu Lingxi;Peng Sihua(Digital and Intelligent Agriculture Research Institute,School of Information Engineering,Huzhou University,Huzhou 313000,China)

机构地区：[1]湖州师范学院信息工程学院数智农业研究院,湖州313000

出　　处：《现代计算机》2024年第20期36-40,共5页Modern Computer

摘　　要：对近年来基于图神经网络的DDI预测方法的研究进展进行了总结,包括DDI预测的定义和分类、基于图神经网络模型的DDI预测方法的步骤、常用的数据集和评价指标以及未来的研究方向和应用前景等。对当前存在的问题,如模型可解释性进行了探讨,并对未来的研究方向例如多模态信息融合等进行了展望。This paper summarizes the research progress of DDI prediction methods based on graph neural network in recent years,including the definition and classification of DDI prediction,the steps of DDI prediction method based on graph neural net-work model,commonly used data sets and evaluation indicators,as well as the future research direction and application prospect.The existing problems,such as model interpretability,are discussed,and the future research directions such as multi-modal infor-mation fusion are prospected.

关 键 词：药物相互作用 图神经网络 深度学习 药物治疗 

分 类 号：R96[医药卫生—药理学] TP183[医药卫生—药学]