Open AccessAnomaly detection of power grid dispatching platform based on Isolation Forest and K-means fusion algorithm
Author(s) -
Sheng Liang,
Kun Zhang,
Wenchong Fang,
Zhifeng Zhou,
Rong Hu,
Wen Zhu,
Yingchen Li,
Yichang Wang,
Jian Hou
Publication year - 2020
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 - 10.1088/1742-6596/1601/2/022010
Subject(s) - isolation (microbiology) , grid , algorithm , computer science , data mining , power grid , power (physics) , anomaly detection , real time computing , mathematics , physics , geometry , quantum mechanics , microbiology and biotechnology , biology
It is very important to detect abnormal Access IP of power grid dispatching platform accurately and rapidly to ensure the safety of power production. An Isolation Forest and K-means fusion algorithm is proposed, which not only solves the deficiency of Isolation Forest only being able to detect anomalies in binary classification, but also solves the defect of Isolation Forest threshold setting based on artificial experience or prior assumptions by designing a threshold setting strategy for Isolation Forest anomalies. The Access IP data of the southern power grid dispatching platform is taken as an example to evaluate the proposed algorithm and model. Through control experiments, the effectiveness and advancedness of the algorithm are illustrated in terms of AUC-ROC curve, accuracy, recall, and F1 score.