摘要
提出基于数据潜在规律挖掘的用户侧窃电行为识别方法。在数据清洗的基础上,利用拉格朗日插值法对缺失数据实施插值填补操作,利用最小-最大标准化方法对插值处理后的用电数据实施标准化处理。然后基于主成分分析设计用户侧用电数据潜在规律挖掘过程,将挖掘结果作为基于L0稀疏超图半监督学习的窃电行为识别方法的识别样本,识别用户用电行为是否存在窃电行为。结果显示:该方法可准确识别用户侧窃电行为,对多个用户用电行为的识别结果符合实际。当正则化参数为0.85时,其对用户侧用电行为的识别结果最可靠。
In order to accurately identify the behavior of stealing electricity from the user side,this study proposes a method of identifying the behavior of stealing electricity from the user side based on data mining.On the basis of data cleaning,Lagrange interpolation method is used to fill the missing data,and minimum-maximum standardization method is used to standardize the electricity consumption data after interpolation.Then,based on principal component analysis,the mining process of the potential rule of user side power consumption data is designed,and the mining results are used as the identification sample of the method based on L0 sparse hypergraph semi-supervised learning to identify whether the user's power consumption behavior is stealing.Experimental results show that this method can accurately identify the behavior of stealing electricity from the user side,and the recognition results of multiple users'behavior are in line with the reality.In the practical application of the method,when the regularization parameter is 0.85,the identification result of user side power consumption behavior is the most reliable.
作者
上官霞
张航
SHANG Guanxia;ZHANG Hang(State Grid Fujian Electric Power Co.,Ltd.,State Grid Fujian Communication Company Fuzhou,350001,China)
出处
《粘接》
CAS
2023年第7期150-154,共5页
Adhesion
