摘要
为了适应Web数据信息高度动态化的发展趋势,针对知识图谱的补全更新实时性欠佳的问题,提出一种知识图谱补全的链接预测方法.该方法运用贝叶斯网络概率推理图模型优势并结合本体推理规则,对知识图谱节点间依赖程度进行定量分析,旨在充分挖掘模型潜在因素,实现正确预测.实验结果表明,该方法能有效提升知识图谱的链接预测效率,保证较高的预测准确度,并及时更新知识图谱.
In order to adapt to the highly dynamic trend of Web data development,and to solve the problem of the real-time performance of knowledge graph updating,a method of knowledge graph link prediction is proposed. This method combined the advantages of Bayesian network probabilistic inference model and the ontology reasoning rules,the confidence among entities is analyzed quantitatively which is aiming at exploiting the potential factors of the model to attain the correct prediction. The results of analysis and experiment show that this method can improve the link prediction efficiency of knowledge graph,ensure the high prediction accuracy,and update the knowledge graph in time.
作者
翟社平
郭琳
高山
段宏宇
李兆兆
马越
ZHAI She-ping;GUO Lin;GAO Shan;DUAN Hong-yu;LI Zhao-zhao;MA Yue(College of Computer Science and Technology, Xi'an University of Posts and Telecommunications, Xi'an 710121, China;College of Economics and Business Administration,Xi'an University of Posts and Telecommunications,Xi'an 710121 ,China})
出处
《小型微型计算机系统》
CSCD
北大核心
2018年第5期995-999,共5页
Journal of Chinese Computer Systems
基金
工业和信息化部通信软科学项目(2017-R-22)资助
陕西省教育厅科学研究计划项目(12JK0733)资助
陕西省社会科学基金项目(2016N008)资助
西安市社会科学规划基金项目(17X63)资助
西安邮电大学研究生创(114-602080105)资助