Abstract:
|
In this talk, we present a new approach for the consistent estimation of the number and location of multiple generalised change-points in noisy data sequences. The number of change-points can increase with the sample size. Examples of signal changes that our method can deal with, are changes in the mean of a piecewise-constant signal and kinks in the continuous, piecewise-linear model. The method is based on an isolation technique, which prevents the consideration of intervals that contain more than one change-point. This isolation enhances the method's accuracy as it allows for detection in the presence of limited spacings between change-points and when the magnitudes of the changes are small. Thresholding and model selection through an information criterion are the two stopping rules described. We will present various examples, in which our method is at least as accurate as the state-of-the-art methods.
|