Exact Multiple Change-Point Detection Via Smallest Valid Partitioning

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Runge, Vincent, Kostic, Anica, Combeau, Alexandre, Romano, Gaetano
Format: Preprint
Published: 2026
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866912875512070144
author Runge, Vincent
Kostic, Anica
Combeau, Alexandre
Romano, Gaetano
author_facet Runge, Vincent
Kostic, Anica
Combeau, Alexandre
Romano, Gaetano
contents We introduce smallest valid partitioning (SVP), a segmentation method for multiple change-point detection in time-series. SVP relies on a local notion of segment validity: a candidate segment is retained only if it passes a user-chosen validity test (e.g., a single change-point test). From the collection of valid segments, we propose a coherent aggregation procedure that constructs a global segmentation which is the exact solution of an optimization problem. Our main contribution is the use of a lexicographic order for the optimization problem that prioritizes parsimony. We analyze the computational complexity of the resulting procedure, which ranges from linear to cubic time depending on the chosen cost and validity functions, the data regime and the number of detected changes. Finally, we assess the quality of SVP through comparisons with standard optimal partitioning algorithms, showing that SVP yields competitive segmentations while explicitly enforcing segment validity. The flexibility of SVP makes it applicable to a broad class of problems; as an illustration, we demonstrate robust change-point detection by encoding robustness in the validity criterion.
format Preprint
id arxiv_https___arxiv_org_abs_2602_04322
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle Exact Multiple Change-Point Detection Via Smallest Valid Partitioning
Runge, Vincent
Kostic, Anica
Combeau, Alexandre
Romano, Gaetano
Methodology
We introduce smallest valid partitioning (SVP), a segmentation method for multiple change-point detection in time-series. SVP relies on a local notion of segment validity: a candidate segment is retained only if it passes a user-chosen validity test (e.g., a single change-point test). From the collection of valid segments, we propose a coherent aggregation procedure that constructs a global segmentation which is the exact solution of an optimization problem. Our main contribution is the use of a lexicographic order for the optimization problem that prioritizes parsimony. We analyze the computational complexity of the resulting procedure, which ranges from linear to cubic time depending on the chosen cost and validity functions, the data regime and the number of detected changes. Finally, we assess the quality of SVP through comparisons with standard optimal partitioning algorithms, showing that SVP yields competitive segmentations while explicitly enforcing segment validity. The flexibility of SVP makes it applicable to a broad class of problems; as an illustration, we demonstrate robust change-point detection by encoding robustness in the validity criterion.
title Exact Multiple Change-Point Detection Via Smallest Valid Partitioning
topic Methodology
url https://arxiv.org/abs/2602.04322