摘要
为了灵敏、准确地识别综合能源异常运行状态,提出基于关联规则算法的综合能源大数据异常识别方法。由关联规则挖掘模块建立频繁项集,满足最小支持度,将这些频繁项集相互连接以生成候选项集,剔除不频繁的候选集。利用最小可信度阈值制定关联规则,并将综合能源大数据分成多种关联规则,利用模糊C均值聚类算法关联规则之间的隶属度,迭代计算后获取最小的价值函数,输出综合能源大数据异常识别结果。实验结果表明,所提出的方法能保障综合能源大数据的完整性,准确、高效地完成数据异常识别。
In order to recognize the anomaly operation state of comprehensive energy sensitively and accurately,an anomaly recognition method of comprehensive energy big data based on association rule algorithm is proposed.The association rule mining module establishes frequent itemsets that meets the minimum support threshold.These frequent itemsets are then connected to themselves to generate candidate itemsets,and infrequent candidate sets are eliminated.Using the minimum confidence threshold,association rules are formulated,and the comprehensive energy big data are divided into multiple association rules.The membership degree between fuzzy C-means clustering algorithm association rules is used to obtain the minimum value function after iterative calculation,and the anomaly recognition results of comprehensive energy big data are output.The experimental results show that the proposed method can ensure the integrity of the comprehensive energy big data,and accurately and efficiently complete data anomaly recognition.
作者
李方军
高建勇
杨海龙
申富泰
王琼
周永博
LI Fangjun;GAO Jianyong;YANG Hailong;SHEN Futai;WANG Qiong;ZHOU Yongbo(State Grid Gansu Electric Power Company,Lanzhou 730050,China;SGIT-UNI Cloud Data Technology Co.,Ltd.,Lanzhou 730000,China)
出处
《微型电脑应用》
2025年第12期260-264,共5页
Microcomputer Applications
关键词
综合能源大数据
关联规则算法
模糊C均值聚类算法
异常识别
价值函数
comprehensive energy big data
association rule algorithm
fuzzy C-means clustering algorithm
anomaly recognition
value function
