期刊文献+

融合表示学习的文档链接网络语义社区发现

Semantic Community Discovery in Document Link Network Based on Representation Learning
下载PDF
导出
摘要 大数据时代背景下诸如论文之间的引用网络、万维网、微博用户之间的网络比比皆是,通常建模为链接网络。利用这些网络链接和节点属性实现语义社区发现有助于了解网络的语义信息和中观结构。现有的语义社区发现方法可利用链接和数学实现语义社区发现,但没有结合基于低维嵌入空间的表示;现有考虑低维嵌入的社区发现方法可更准确挖掘社区结构,但没有利用文档的内容属性。这些方法都不能充分利用链接网络的细粒度结构和语义信息,提出一种融合网络节点表示学习的属性网络的语义社区发现模型Rcolc(Representation learning and Community discovery on links and contents)。该模型可以融合文档的链接和属性信息实现语义社区发现,并考虑文档的基于链接的低维嵌入提升社区发现准确性。在真实属性网络上的实验表明该算法优于主流算法。 In the era of big data,there are many networks,such as citation network between papers,world wide web and microblog users,which are usually modeled as link network.Using these network links and node attributes to realize semantic community discovery helps to understand the semantic information and meso structure of the network.The existing semantic community discovery methods can use links and mathematics to realize semantic community discovery,but they do not combine the representation based on low dimensional embedded space;The existing community discovery methods considering low dimensional embedding can mine community structure more accurately,but they do not use the content attributes of documents.None of these methods can make full use of the fine-grained structure and semantic information of link networks.This paper proposes a semantic community discovery model RColc(Representation learning and Community discovery on links and contents)based on attribute networks with network node representation learning.The model can integrate link and attribute information of documents to realize semantic community discovery,and consider low dimensional embedding of documents based on link to improve the accuracy of community discovery.Experiments on real attribute networks show that the algorithm is superior to the mainstream algorithms.
作者 郭江林 GUO Jiang-lin(School of College of Information Engineering,Hebei GEO University,Shijiazhuang 050031,China)
出处 《新一代信息技术》 2021年第10期37-41,共5页 New Generation of Information Technology
基金 河北省高等学校科学技术研究项目(项目编号:ZD2020175)。
关键词 语义社区发现 网络嵌入 节点表示 semantic community discovery network embedding node representation
  • 相关文献

相关作者

内容加载中请稍等...

相关机构

内容加载中请稍等...

相关主题

内容加载中请稍等...

浏览历史

内容加载中请稍等...
;
使用帮助 返回顶部