Open AccessResearch on An Ensemble Anomaly Detection Algorithm
Author(s) -
Yupeng Wang,
Shibing Zhu,
Changqing Li
Publication year - 2019
Publication title -
journal of physics. conference series
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.21
H-Index - 85
eISSN - 1742-6596
pISSN - 1742-6588
DOI - 10.1088/1742-6596/1314/1/012198
Subject(s) - anomaly detection , computer science , algorithm , anomaly (physics) , process (computing) , ensemble learning , majority rule , data mining , artificial intelligence , physics , condensed matter physics , operating system
Aiming at the problem that the applicability of single anomaly detection algorithm is not strong in aerospace experiment, an ensemble anomaly detection algorithm is proposed. This algorithm combines multiple machine algorithms and can obtain better detection performance than any other algorithm. Through comparison, k-NN, PCA and HBOS are selected. These three algorithms have fast calculation speed and different algorithm mechanisms, which can effectively process various data sets. This paper first introduces the basic concept of anomaly detection, then introduces and explains the three algorithms, then integrates the three algorithms, and introduces the voting mechanism to vote on whether the sample points are normal. Finally, the performance of the algorithm is tested through simulation experiments. Compared with a single algorithm, the ensemble algorithm has better performance in precision and accuracy.