Multiple change-point detection via bottom-up scanning
Abstract
We study nonparametric multiple change-point detection for high-dimensional sequences, aiming to identify time points at which the underlying distribution changes. While many existing methods perform well when change-points are well-separated, their performance can deteriorate when structural breaks are densely clustered. To address this challenge, we propose gBottomup, a graph-based bottom-up framework for multiple change-point detection in high-dimensional settings. gBottomup constructs a hierarchical segmentation by proposing merges of adjacent segments and verifying them through an unmerge rule that combines absolute significance with relative local heterogeneity, thereby adaptively refining partitions and estimating the number of change-points. Simulation results demonstrate that gBottomup performs reliably across a range of structural configurations and is particularly effective in frequent change-point settings, where existing top-down procedures may lose sensitivity. Runtime experiments indicate favorable computational performance relative to graph-based top-down alternatives. We illustrate the proposed method through an analysis of a S&P 500 dataset.
Related
- Multiple change-point detection for Poisson processes
- Laplacian U-Processes for Multiple Change-Point Detection in Dependent Text Networks: An Application to Historical Chinese Articles
- Multiple Change Point Detection in the Mean and Covariance of High-Dimensional Matrix Time Series
- Exact Multiple Change-Point Detection Via Smallest Valid Partitioning
- The Sample Complexity of Multiple Change Point Identification under Bandit Feedback
BibTeX
@misc{park2026multiple,
title = {{Multiple change-point detection via bottom-up scanning}},
author = {Jinhyeok Park and Hyeyoung Maeng and Hoseung Song},
howpublished = {arXiv},
year = {2026},
month = sep,
doi = {10.48550/arxiv.2609.31116},
eprint = {2609.31116},
archivePrefix = {arXiv},
url = {https://arxiv.org/abs/2609.31116},
note = {CC BY 4.0},
}
RIS
TY - UNPB
TI - Multiple change-point detection via bottom-up scanning
AU - Jinhyeok Park
AU - Hyeyoung Maeng
AU - Hoseung Song
PY - 2026
DA - 2026/09/25
PB - arXiv
DO - 10.48550/arxiv.2609.31116
UR - https://arxiv.org/abs/2609.31116
AB - We study nonparametric multiple change-point detection for high-dimensional sequences, aiming to identify time points at which the underlying distribution changes. While many existing methods perform well when change-points are well-separated, their performance can deteriorate when structural breaks are densely clustered. To address this challenge, we propose gBottomup, a graph-based bottom-up framework for multiple change-point detection in high-dimensional settings. gBottomup constructs a hierarchical segmentation by proposing merges of adjacent segments and verifying them through an unmerge rule that combines absolute significance with relative local heterogeneity, thereby adaptively refining partitions and estimating the number of change-points. Simulation results demonstrate that gBottomup performs reliably across a range of structural configurations and is particularly effective in frequent change-point settings, where existing top-down procedures may lose sensitivity. Runtime experiments indicate favorable computational performance relative to graph-based top-down alternatives. We illustrate the proposed method through an analysis of a S&P 500 dataset.
ER -