基于知识图谱和深度学习的学术论文推荐算法研究
Research on academic paper recommendation algorithm based on knowledge mapping and deep learning
杨宇亮 1石嘉豪 1陶秀杰 1林杰1
作者信息
- 1. 南方电网科学研究院有限责任公司,广东广州 510000
- 折叠
摘要
目的:文章针对当前学术论文推荐中存在的不足,提出将知识图谱与深度学习相结合的研究思路,研究一种学术论文推荐算法.方法:搜集文献资料,建立相关知识图谱.对用户的行为喜好及检索要求进行了分析.方法:通过构造查询矢量,度量相似度,生成资源推荐表,完成学术论文推荐.结论:通过与传统算法的比较,本文提出的算法可以使用户的搜索准确率提高4.53％,并且能够显著地提高用户的搜索准确率.
Abstract
Purpose:This article aims at the shortcomings in current academic paper recommenda-tion,proposes a research idea that combines knowledge graphs with deep learning,and studies an academic paper recommendation algorithm.Methods:Collect literature and establish relevant knowledge graph.The user's behavioral preferences and retrieval requirements were analyzed.Method:Complete the recommendation of academic papers by constructing query vectors,meas-uring similarity,and generating resource recommendation tables.Conclusion:By comparing with the traditional algorithm,the algorithm proposed in this article can increase the user's search accuracy by 4.53％,and can significantly improve the user's search accuracy.
关键词
知识图谱/深度学习/学术论文Key words
knowledge graph/deep learning/academic paper引用本文复制引用
出版年
2024
NETL
NSTL
万方数据