辅以大众分类法的多标签分类深度学习信息架构  

Multi-label Classification Deep Learning Information Architecture Supplemented by Folksonomy

作　　者：张静怡 王宇晓 ZHANG Jingyi;WANG Yuxiao(China Academy of Information and Communications Technology,Beijing 100191,China;Key Laboratory of Mobile Application Innovation and Governance Technology,Ministry of Industry and Information Technology,Beijing 100191,China)

机构地区：[1]中国信息通信研究院,北京100191 [2]移动应用创新与治理技术工业和信息化部重点实验室,北京100191

出　　处：《现代信息科技》2025年第1期104-109,共6页Modern Information Technology

基　　金：工业和信息化部2022年产业技术基础公共服务平台(2022-234-226)。

摘　　要：随着移动互联网的迅速发展,信息量和信息维度急速增长,如何组建起便于用户查找和理解信息的信息架构成为亟待优化的问题。文章重点探讨信息分类技术中的大众分类法对信息架构的积极作用,在此基础上,利用深度学习算法在处理多标签分类问题时可对多标签间的相关性和依赖关系进行自动建模的特点,提出一种辅以大众分类算法的多标签分类深度学习信息架构,提高了信息分类关联的准确度和相关度,为用户提供了便利。With the rapid development of the mobile Internet,the amount of information and information dimensions are growing rapidly.How to set up an information architecture that is convenient for users tofind and understand information has become an urgent problem to be optimized.This paper focuses on the positive effect of folksonomy in information classification technology on information architecture.On this basis,this paper uses the characteristics that the Deep Learning algorithm can automatically model the correlation and dependency relationships among multi-labels when dealing with multi-label classification problems,to propose a multi-label classification Deep Learning information architecture supplemented by folksonomy algorithm.It improves the accuracy and relevance of information classification association and provides convenience for users.

关 键 词：大众分类 深度学习 信息架构 多标签分类 

分 类 号：TP18[自动化与计算机技术—控制理论与控制工程]