Multiple Change Points Detection Method Based on TSTKS and CPI Sliding Window Strategy

摘要

This paper takes time series data as the research object, proposes a TSTKS (Trigeminal search tree and KS Statistic) change point detection method based on CPI (Change Point Interval) sliding window strategy. First of all, this paper constructs a TSTKS change point detection method, which improves the accuracy of change point detection. Secondly, for the characteristics of multiple change points for time series data. Based on the fixed sliding window model, a CPI sliding window strategy is proposed, conditionally change the window size of the sliding window. It improves the accuracy of change detection and shortens the detection time. Finally, in order to verify the effect of the TSTKS change point detection method based on CPI sliding window strategy, simulation experiments were performed and used to detect the change points of epilepsy brain wave data. The results show the effectiveness of the proposed method and has practical application value.