Open AccessAnomaly Detection Method for Operation and Maintenance Data Based on One-Class Learning
Author(s) -
Ji’an Duan,
Bo Yang,
Xiang Wei,
Ning Liu,
Yintie Zhang
Publication year - 2020
Publication title -
iop conference series. earth and environmental science
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.179
H-Index - 26
eISSN - 1755-1307
pISSN - 1755-1315
DOI - 10.1088/1755-1315/440/3/032146
Subject(s) - anomaly detection , computer science , sample (material) , class (philosophy) , data mining , anomaly (physics) , fault (geology) , support vector machine , artificial intelligence , chemistry , physics , chromatography , seismology , geology , condensed matter physics
Due to frequent changes in system services, the amount of data collected in a certain state is insufficient, which causes problems such as insufficiency of normal sample data, scarcity of fault sample data, and lack of prior knowledge. Aiming at the problem of anomaly detection of small sample operation and maintenance data lacking negative samples, this paper proposes an operation and maintenance data anomaly detection method based on one-class learning, which uses SVDD (support vector data description) method to eliminate abnormal data in the collected operation and maintenance data. Then, we can better analyze the subsequent data. The experiments show that the proposed method is reasonable and effective.