面向多源异构数据的语义对象识别技术研究  

Research on Semantic Object Recognition Technology for Multi-source Heterogeneous Data

作　　者：萧展辉 柴雁欣 李学龄 包新晔 XIAO Zhan-hui;CHAI Yan-xin;LI Xue-ling;BAO Xin-ye(China Southern Power Grid Digital Power Grid Research Institute Co.,Ltd.,Guangzhou 510700 China)

机构地区：[1]南方电网数字电网研究院有限公司,广东广州510700

出　　处：《自动化技术与应用》2024年第12期103-106,共4页Techniques of Automation and Applications

基　　金：广东电网有限责任公司数字电网关键技术(国重配套)(JY-KF-03-GX-21-016)。

摘　　要：为了更好地应用电力系统中的多源异构数据,优化电力系统决策环节,提高数据挖掘分析能力,提出面向多源异构数据的语义对象识别技术。构建多源异构数据电网数字模型,并使用核函数实现数据融合。使用最大熵马尔科夫模型构建电网信息知识图谱。根据SDK对外语义服务过程,使用周期模式挖掘算法挖掘数据,并获取语义对象轨迹,完成语义识别。构建实验分析环节,应用3种数据库验证此技术应用效果。实验结果表明:此技术具有较高的数据分析能力,可在一定程度上提升语义对象识别准确性。In order to better apply the multi-source heterogeneous data in power system,optimize the decision-making process of power system,and improve the ability of data mining and analysis,a semantic object recognition technology for multi-source heterogeneous data is proposed.It builds a multi-source heterogeneous data grid digital model,and uses kernel function to realize data fusion.The Maximum Entropy Markov model is used to construct the knowledge map of power grid information.According to the SDK external semantic service process,the periodic pattern mining algorithm is used to mine data,obtain semantic object trajectories,and complete semantic recognition.It constructs the experimental analysis link,and applies three kinds of databases to verify the application effect of this technology.The experimental results show that this technology has high data analysis ability and can improve the accuracy of semantic object recognition to a certain extent.

关 键 词：多源数据融合 语义识别 多源异构数据 知识图谱 信息空间维度 模型 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]