基于机器学习的实时云数据关联规则提取与运维分析方法  

Real time cloud data association rule extraction and operation and maintenance analysis method based on machine learning

作　　者：倪家明[1] 张耀 刘鹏 贺小刚 王丹丹 NI Jiaming;ZHANG Yao;LIU Peng;HE Xiaogang;WANG Dandan(State Grid Tianjin Electric Power Company,Tianjin 300010,China;Tianjin Sanyuan Electric Information Technology Co.,Ltd.,Tianjin 300010,China)

机构地区：[1]国网天津市电力公司,天津300010 [2]天津三源电力信息技术股份有限公司,天津300010

出　　处：《电子设计工程》2022年第4期125-128,133,共5页Electronic Design Engineering

摘　　要：以往的数据关联规则提取方法的项集标记结果不全面。针对该问题,提出了基于机器学习的实时云数据关联规则提取方法,为数据运维分析奠定良好的基础。构建机器学习网络架构,利用无监督训练和调优两个步骤训练机器学习网络,由此构建任务调度模型,以最小化构建排队处理提取与运维任务。依据机器学习扫描原理寻找强项集,从强度集合中导出关联规则。在此基础上,描述关联规则运维流程,删除不符合最低支持度的项目集。实验结果表明,该方法获取的实例数据集与实际情况一致,说明其提取效果较好。The previous data association rule extraction methods do not have comprehensive itemset labeling results.To solve this problem,a real-time cloud data association rule extraction method based on machine learning is proposed,which lays a good foundation for data operation and maintenance analysis. The machine learning network architecture is constructed,and the machine learning network is trained by unsupervised training and tuning,so as to build a task scheduling model to minimize the construction of queue processing,extraction and operation and maintenance tasks.According to the scanning principle of machine learning,the strength set is found,and the association rules are derived from the strength set. On this basis,describe the operation and maintenance process of association rules and delete the item set that does not meet the minimum support.The experimental results show that the example data set obtained by this method is consistent with the actual situation,which shows that its extraction effect is good.

关 键 词：机器学习 实时云数据 关联规则提取 运维流程 

分 类 号：TN721[电子电信—电路与系统]