Detection of a Sparse Change in High-Dimensional Time Series

TL;DR

This paper introduces a new sparsity likelihood-based method for detecting sparse changes in high-dimensional time series, demonstrating its effectiveness through theoretical guarantees, simulations, and real-world financial data applications.

Contribution

It proposes a novel SL-based change-point detection algorithm for high-dimensional data with general covariance, with proven asymptotic error probability convergence.

Findings

SL-based method achieves asymptotic error probability convergence

Simulation studies confirm the method's effectiveness

Application to S&P500 data demonstrates practical utility

Abstract

Consider the detection of a sparse change in high-dimensional time-series. We introduce Sparsity Likelihood-based (SL-based) score and the change-points detection procedure in multivariate normal model with general covariance structure. SL-based algorithm is proved to achieve that supremum of error probabilities converges to 0. We run the simulation studies for SL-based algorithm and also illustrate its applications to a S&P500 dataset.