Change Point Detection Via Sub-Gaussian Fitting.
Adam W. Hoover, Anirud Singh, Stephanie R. Fishel, Eric R. Muth · 2010
Abstract This work presents a novel approach to detect a change in the state of a signal. We propose that in a given state, the values of a signal vary in a subrange of a Gaussian distribution. We describe methods to monitor a signal in real time for change points based upon sub-Gaussian fitting. The proposed algorithm was implemented and tested on heartrate variability data. signal state change Key words: change point detection, state change, CUSUM 1