Multiple-change-point detection for high dimensional time series via sparsified binary segmentation

Abstract

Time series segmentation, a.k.a. multiple change-point detection, is a well-established problem. However, few solutions are designed specifically for high-dimensional situations. In this paper, our interest is in segmenting the second-order structure of a high-dimensional time series. In a generic step of a binary segmentation algorithm for multivariate time series, one natural solution is to combine CUSUM statistics obtained from local periodograms and cross-periodograms of the components of the input time series. However, the standard "maximum" and "average" methods for doing so often fail in high dimensions when, for example, the change-points are sparse across the panel or the CUSUM statistics are spuriously large. In this paper, we propose the Sparsified Binary Segmentation (SBS) algorithm which aggregates the CUSUM statistics by adding only those that pass a certain threshold. This "sparsifying" step reduces the impact of irrelevant, noisy contributions, which is particularly beneficial in high dimensions. In order to show the consistency of SBS, we introduce the multivariate Locally Stationary Wavelet model for time series, which is a separate contribution of this work.

References in corpus (3)

• Wild binary segmentation for multiple change-point detection
• Multiscale and multilevel technique for consistent segmentation of nonstationary time series
• Multiscale interpretation of taut string estimation and its connection to Unbalanced Haar wavelets

Cited by in corpus (8)

• A test for second-order stationarity of time series based on unsystematic sub-samples
• Latent Neural Stochastic Differential Equations for Change Point Detection
• Estimation of dynamic networks for high-dimensional nonstationary time series
• Selective linear segmentation for detecting relevant parameter changes
• A Fast Detection Method of Break Points in Effective Connectivity Networks
• Translation-invariant functional clustering on COVID-19 deaths adjusted on population risk factors
• Data-adaptive structural change-point detection via isolation
• Spectral Change Point Estimation for High Dimensional Time Series by Sparse Tensor Decomposition