An Improved Data Anomaly Detection Method Based on Isolation Forest

Улучшенный метод обнаружения аномалий в данных на основе изолирующего леса
Dong Ling Xu, Yanjun Wang, Yulong Meng, Ziying Zhang
2017-12-01

Isolation ForestLocal Outlier FactorSA-iForestdata anomaly detectionsimulated annealing
An improved data anomaly detection method SA-iForest is proposed to solve the problem of low accuracy, poor execution efficiency and generalization ability of data anomalies detection algorithm based on isolated forest. Based on the idea of selective integration, the precision and the difference value are taken as the criterion, and the simulated annealing algorithm is used to select the isolation tree with high abnormality detection and differently to optimize the forest. At the same time, the excess detection precision is small and the difference is small Isolation tree improves the forest construction process of isolated forests, which improves the efficiency of the algorithm and improves the efficiency of the algorithm. The method of data anomaly detection based on SA-iForest is compared with the traditional Isolation Forest algorithm and LOF algorithm, and the accuracy and efficiency of the data are verified by the standard simulation data set. There is a significant improvement.
1
A simulated annealing algorithm identifies isolation trees with strong anomaly-detection performance and complementary behavior to optimize the forest.
2
Experiments on standard simulated datasets show significant improvements in anomaly-detection accuracy and efficiency over traditional Isolation Forest and LOF.
3
Removing isolation trees with low detection precision and low diversity improves forest construction efficiency.
4
SA-iForest improves Isolation Forest anomaly detection by selectively integrating isolation trees using detection precision and diversity as selection criteria.

data anomaly detection based on isolation forests

detection accuracy, execution efficiency, and generalization ability improved through selective isolation-tree ensemble optimization

Publication Details
Publication Date
2017-12-01
Journal
Publisher
ISSN
Cited by
156
Access Type
Author Information
Authors
Dong Ling Xu
Yanjun Wang
Yulong Meng
Ziying Zhang
Explore further
Open the scid.ai AI chat with a ready-made request: it will find papers on a similar topic and help build a literature review.
Find similar papers in the chat →
Make a presentation
100%