High-dimensional data analysis: Change point detection via bootstrap MOSUM

Abstract

Change point detection in high-dimensional data has become a significant area of research in the era of big data. In this paper, we propose a novel test statistic for high-dimensional change point detection based on the bootstrap moving sum (MOSUM) method. We derive the theoretical properties of the proposed statistic and establish the consistency of the change point location estimator. Numerical simulation results demonstrate that our method outperforms the bootstrap cumulative sum (CUSUM) test statistic. Finally, we apply the proposed method to empirically analyze a real-world data set.