Studies Of The Technology Of Anomaly Detection Based On Data Streams And Its Application

Recent years, the technology of anomaly detection in data streams has been attracting more and more attention due to its broad potential applications in venture analysis, network monitoring, trend analysis and so on. The technology has been studied, and the rational model of anomaly detection has been established using the technology of sliding window and the anomaly detection algorithm of shifted wavelet tree(SWT). Aiming at the limitation of the algorithm of SWT, an improved algorithm was proposed: ratio threshold was used to enhance the accuracy of detection; the binary detection was used to reduce response time; the algorithm of real-time incremental update was used to meet the online processing; the improved algorithm of setting threshold was used to reduce the misreport rate. Anomaly detection simulation was implemented on Gamma Ray data and Power Quality Disturbance data using the improved algorithm, the results have been analyzed and appraised completely, and a comparison with the algorithm of SWT and error tree has been made to show the effectiveness of the proposed algorithm.