电商领域多模态商品知识图谱构建研究  

Research on construction of multimodal commodity knowledge graph in e⁃commerce domain

作　　者：宿恺[1] 潘晨辉 SU Kai;PAN Chenhui(School of Management,Shenyang University of Technology,Shenyang 110003,China)

机构地区：[1]沈阳工业大学管理学院,辽宁沈阳110003

出　　处：《现代电子技术》2023年第20期173-177,共5页Modern Electronics Technique

基　　金：辽宁省科技厅重点研发项目(2019JH8/10100068)。

摘　　要：在电子商务蓬勃发展的背景下,电子商务平台的推荐系统面临着冷启动和数据稀疏性的问题。而构建多模态知识图谱可以为电子商务平台实际应用提供重要的支撑。为此,首先指出当前电子商务中推荐系统的难点和传统知识图谱模态缺失的情况,设计多模态知识图谱构建的总体框架,分析多模态数据来源;然后,论述电商领域多模态商品知识图谱构建过程中的核心技术;最后,举例说明多模态商品知识图谱在电子商务平台中的实际应用。结果表明,多模态商品知识图谱可为电子商务的发展提供支持,未来多模态商品知识图谱的发展需要以实际应用为导向和多领域融合为方法,推动多模态商品知识图谱的优化发展。In the backdrop of the booming development of e⁃commerce,recommendation systems on e⁃commerce platforms face challenges of cold start and data sparsity.Building a multimodal knowledge graph can provide important support for the practical application of e⁃commerce platforms.The difficulties of current recommendation systems in e⁃commerce and the lack of traditional knowledge graph modes are pointed out,the overall framework is designed for constructing multimodal knowledge graphs,and the sources of multimodal data is analyzed.The core technologies in the construction process of multimodal commodity knowledge graph in the e⁃commerce field are discussed,and examples of the multimodal product knowledge graph in e⁃commerce platforms are given to show the practical application.The results indicate that the multimodal commodity knowledge graph can provide support for the development of e⁃commerce.In the future,the development of multimodal commodity knowledge graph needs to be guided by practical applications and integrated into multiple fields to promote the optimization and development of multimodal commodity knowledge graph.

关 键 词：电子商务 多模态数据 知识图谱构建 信息抽取 知识表示 知识融合 图谱存储 

分 类 号：TN919-34[电子电信—通信与信息系统] TP311[电子电信—信息与通信工程]