Changepoint detection with the use of the RESPERM method – a Monte Carlo study

Abstract

RESPERM (residuals permutation-based method) is a single changepoint detection method based on regression residuals permutation, which can be applied to many physiological situations where the regression slope can change suddenly at a given point. This article presents the results of a Monte Carlo study on the properties of the RESPERM method for single changepoint detection in a linear regression model. We compared our method with a well-known segmented method for detection breakpoint in linear models. The Monte Carlo study showed that when the input data are very noisy, the RESPERM method outperforms the segmented approach in terms of variance, and in the case of bias, the results of the two methods are comparable.