A Fast Method of Anomaly Detection based on Adaptive Slide Window (ASW) and Wave Vector Classification (WVC)
Qing Ren, QI Jin-peng, Zhong JinMei, Yitong Cao, Zhu JunJun · 2021 17th International Conference on Computational Intelligence and Security (CIS) · 2021
Traditional data analysis is difficult to be fast and accurate for large-scale time-series data. To solve the problems encountered in data analysis, this paper combines the wave vector classification (WVC) technology with the strategy of adaptive slide window (ASW) and gives a multiple change points (MCPs) method for lesion time series data. Based on the TSTKS change point detection algorithm and combined with the sliding window theory, this method calculates the fluctuation vectors of all windows for multi-layer classification, feeds back the window detection results to the next window, and then changes the window size to realize the state analysis and rapid diagnosis of online sequential pathological data. Simulation experiments and pathological data analysis results show that the new method proposed in this paper is faster and more efficient, and can be used as a new online change point detection method for time series data.