基于图注意力机制和Transformer的异常检测
Abnormal Detection Based on Graph Attention Mechanisms and Transformer
查看参考文献30篇
|
文摘
|
异常检测对电力行业的发展有着重要的影响,如何根据大规模电力数据进行异常检测是重要的研究热点.目前,大多数研究通过聚类或神经网络进行异常检测.但是这些方法忽略了时序数据之间潜在的关联关系及某些特点的重要信息,没有充分挖掘出数据的潜在价值.因此,提出了一种基于图注意力和Transformer的异常检测模型.该模型首先根据数据中台中获取的电力数据(主要包括用户ID、电能表ID、用户类型、电流、电压、功率等数据)构建一个异构信息网络;然后,为了减少模型参数和避免出现过拟合的现象,在图卷积网络(Graph Convolutional Network, GCN)模型的基础上,引入非负矩阵分解(Non-Negative Matrix Factorization,NNMF)的方法来进行相似性学习;最后采用图注意力网络(Graph Attention Network,GAT)和Transformer共同捕获数据间的相互关联关系,从而提高检测精度.以中国某地区的电力数据为基础进行验证,实验结果表明所提出的方法可以有效进行异常检测. |
|
其他语种文摘
|
Anomaly detection has an important impact on the development of the electric power industry, and how to detect anomalies based on large-scale power data is a research hotspot. At present, most researches use clustering or neural network to detect anomalies. But these methods ignore the potential relationship between the data and miss some specific important information, and do not fully exploit the potential value of the data. Therefore, an abnormal detection model based on graph attention and transformer is proposed. The model first constructs a heterogeneous information network based on the power data(mainly including user ID, meter ID, user type, electrical current, voltage, power, etc.)collected in the data center; then, in order to reduce the model parameters and avoid the phenomenon of overfitting, on the basis of the graph convolutional network(GCN)model, a non-negative matrix factorization(NNMF)method is introduced to perform similarity learning; finally, a graph attention network(GAT)and Transformer are jointly used to capture the correlation relationships between data, thus improving the detection accuracy. The validation analysis is carried out based on the power data of a region in China. The experimental results show that the proposed method can effectively perform anomaly detection. |
|
来源
|
电子学报
,2022,50(4):900-908 【核心库】
|
|
DOI
|
10.12263/DZXB.20210722
|
|
关键词
|
异常检测
;
异构信息网络
;
相似性学习
;
图注意力网络
;
Transformer
|
|
地址
|
1.
国网山东省电力公司信息通信公司, 山东, 济南, 250013
2.
山东大学软件学院, 山东, 济南, 250101
|
|
语种
|
中文 |
|
文献类型
|
研究性论文 |
|
ISSN
|
0372-2112 |
|
学科
|
自动化技术、计算机技术 |
|
基金
|
国网山东省电力公司科技项目
|
|
文献收藏号
|
CSCD:7190616
|
参考文献 共
30
共2页
|
|
1.
谢敬东. 基于分阶段离群点检测的电力市场异常辨识.
科学技术与工程,2021,21(9):3633-3641
|
CSCD被引
2
次
|
|
|
|
|
2.
Al-Dhamari A. Online video-based abnormal detection using highly motion techniques and statistical measures.
Telkomnika,2019,17(4):2039-2047
|
CSCD被引
1
次
|
|
|
|
|
3.
丁小欧. 基于相关性分析的工业时序数据异常检测.
软件学报,2020,31(3):22
|
CSCD被引
2
次
|
|
|
|
|
4.
Lim J. Web based online real-time outage cost assessment information system of power system.
Review of Scientific Instruments,2012,37(2):171-172
|
CSCD被引
1
次
|
|
|
|
|
5.
Braitman L E. Confidence intervals extract clinically useful information from data.
Annals of Internal Medicine,1988,108(2):296-298
|
CSCD被引
1
次
|
|
|
|
|
6.
Zhang S. Design of wide-area power system damping controllers resilient to communication failures.
IEEE Transactions on Power Systems,2013,28(4):4292-4300
|
CSCD被引
6
次
|
|
|
|
|
7.
Shafaghi A. Equipment failure rate updating-Bayesian estimation.
Journal of Hazardous Materials,2008,159(1):87-91
|
CSCD被引
2
次
|
|
|
|
|
8.
Sorensen P. Power fluctuations from large wind farms.
IEEE Transactions on Power Systems,2007,22(3):958-965
|
CSCD被引
40
次
|
|
|
|
|
9.
Xu L. Power distribution fault cause identification with imbalanced data using the data mining-based fuzzy classification E-algorithm.
IEEE Transactions on Power Systems,2007,22(1):164-171
|
CSCD被引
7
次
|
|
|
|
|
10.
Bu S. When the smart grid meets energy Efficient communications: Green wireless cellular networks powered by the smart grid.
IEEE Transactions on Wireless Communications,2012,11(8):3014-3024
|
CSCD被引
5
次
|
|
|
|
|
11.
孙毅. 基于高斯核函数改进的电力用户用电数据离群点检测方法.
电网技术,2018,42(5):1595-1606
|
CSCD被引
25
次
|
|
|
|
|
12.
Monedero I. Detection of frauds and other non-technical losses in a power utility using Pearson coeffificient, Bayesian networks and decision trees.
International Journal of Electrical Power & Energy Systems,2012,34(1):90-98
|
CSCD被引
13
次
|
|
|
|
|
13.
Wang Z. Analysis of 10kV nontechnical loss detection with data-driven approaches.
2019 IEEE Innovative Smart Grid Technologies-Asia (ISGT Asia),2019:4154-4158
|
CSCD被引
1
次
|
|
|
|
|
14.
Buzau M M. Hybrid deep neural networks for detection of non-technical losses in electricity smart meters.
IEEE Transactions on Power Systems,2020,35(2):1254-1263
|
CSCD被引
10
次
|
|
|
|
|
15.
Chahla C. A deep learning approach for anomaly detection and prediction in power consumption data.
Energy Efficiency,2020,13(8):1633-1651
|
CSCD被引
2
次
|
|
|
|
|
16.
Barua A. Hierarchical temporal memory based machine learning for real-time, unsupervised anomaly detection in smart grid: WiP abstract.
2020 ACM/IEEE 11th International Conference on Cyber-Physical Systems(ICCPS),2020:188-189
|
CSCD被引
1
次
|
|
|
|
|
17.
Rouzbahani H M. An ensemble deep convolutional neural network model for electricity theft detection in smart grids.
2020 IEEE International Conference on Systems, Man, and Cybernetics(SMC). Singapore,2020:3637-3642
|
CSCD被引
1
次
|
|
|
|
|
18.
Lo Y L. Non-technical loss detection using smart distribution network measurement data.
IEEE PES Innovative Smart Grid Technologies,2012:1-5
|
CSCD被引
4
次
|
|
|
|
|
19.
张承智. 基于实值深度置信网络的用户侧窃电行为检测.
电网技术,2019,43(3):1083-1091
|
CSCD被引
28
次
|
|
|
|
|
20.
Wang B Q. Variance error of multi-classification based anomaly detection for time series data.
J. Comput. Methods Sci. Eng,2021,21(4):875-890
|
CSCD被引
1
次
|
|
|
|
|
了解气象卫星更多信息，请访问