基于共振思想的变点检测及形成机理研究OACHSSCD

Due to the fact that data change points are characterized by difficulty to be fully observed and also have extremely complex generation mechanisms,there exist many difficulties in detecting change points.In view of this,the paper takes the loga-rithmic rate of return of stock market index and commodity market index as research objects,depends on the resonance theory of physics,uses Complementary Ensemble Empirical Mode Decomposition to analyze the transient characteristic parameters of the main influencing factors,and determine the resonance factors.Finally,based on resonance factors,the paper designs a frequency domain detection method for data change points to explore the mechanism of change point formation.The research findings are as follows:The transient characteristic parameters of different influencing factors have specific relationships at the time of change point formation,which can be used to detect change points.Moreover,it can be considered that the resonance effect between the corresponding components of the system is the main cause of the formation of change points.