Nonparametric method of structural break detection in stochastic time series regression model

We propose a novel nonparametric test to detect structural breaks in the conditional mean and/or variance of a time series.

Our method does not assume any specific parametric form for the dependence structure of the regressor, the time series model, or the distribution of the noise.

This flexibility allows our algorithm to be applicable to a wide range of framework.

We further apply the proposed test to accurately localize the changepoints and establish theoretical guarantees showing that the estimated structural breaks are consistent, meaning they lie sufficiently close to the true breakpoints when a sufficiently large sample is available.

The effectiveness of the proposed algorithm is demonstrated through an extensive simulation study encompassing a diverse range of time series structures, including light, moderately heavy, and heavy tailed distributions.

We also show a real-life example, where an application to Bitcoin prices and Google search volume illustrates how the procedure can identify changes in the conditional relationship between market attention and price dynamics.