A new A new algorithm for time series change point and its application on deformation analysis
Wang Hong · Cehui kexue · 2009
In this paper,a new algorithm is introduced for change point of time series based on self-adaptive artificial neural network.Firstly,the length of moving window d and interval of moving n0 are chosen properly,and original time series can be transformed into new series.Secondly,new series are taken as input of network and divided into two types.Then the number and location of change point can be obtained.Finally,the kinds of change point are estimated.Simulative data and empirical data are tested for effectiveness of the algorithm.