High dimensional time series anomaly detection based on generated adversarial network and spatial correlation

Junru Pan; Ying Liang; Zexin Chen

doi:10.1117/12.2675302

25 May 2023 High dimensional time series anomaly detection based on generated adversarial network and spatial correlation

Junru Pan, Ying Liang, Zexin Chen

Author Affiliations +

Proceedings Volume 12636, Third International Conference on Machine Learning and Computer Application (ICMLCA 2022); 1263652 (2023) https://doi.org/10.1117/12.2675302
Event: Third International Conference on Machine Learning and Computer Application (ICMLCA 2022), 2022, Shenyang, China

Abstract

The production data collected by a large number of sensors in actual industrial production have multiple sources, high collection frequency and multiple dimensions, and the existing cleaning methods are difficult to accurately model the changeable time series. Therefore, a high dimensional time series cleaning method based on generative adversarial network and spatial correlation is proposed. Firstly, the time series with multiple dimensions are piecewise aggregated data processing. Secondly, the correlation analysis is carried out by mathematical statistical method, and the weighted repair is made after the abnormal value is found in the sequence. Finally, the long and short term memory cycling neural network training data set was used to complete the data cleaning, and the accuracy rate, recall rate and F1 score were calculated to evaluate the abnormal detection performance. The experimental results show that the cleaning efficiency and accuracy of this method are significantly improved.

Citation Download Citation

Junru Pan, Ying Liang, and Zexin Chen "High dimensional time series anomaly detection based on generated adversarial network and spatial correlation", Proc. SPIE 12636, Third International Conference on Machine Learning and Computer Application (ICMLCA 2022), 1263652 (25 May 2023); https://doi.org/10.1117/12.2675302

ACCESS THE FULL ARTICLE

INSTITUTIONAL
Select your institution to access the SPIE Digital Library.

SELECT YOUR INSTITUTION

PERSONAL
Sign in with your SPIE account to access your personal subscriptions or to use specific features such as save to my library, sign up for alerts, save searches, etc.

PERSONAL SIGN IN

No SPIE Account? Create one

PURCHASE THIS CONTENT

SUBSCRIBE TO DIGITAL LIBRARY

50 downloads per 1-year subscription

Members: $195

Non-members: $335 ADD TO CART

25 downloads per 1 - year subscription

Members: $145

Non-members: $250 ADD TO CART

PURCHASE SINGLE ARTICLE

Includes PDF, HTML & Video, when available