A factor model approach for the joint segmentation with between-series correlation

TL;DR

This paper introduces a factor model-based method for segmenting correlated time-series, utilizing dynamic programming for efficient breakpoint detection and a model selection procedure to determine the number of segments and factors, implemented in an R package.

Contribution

The paper presents a novel approach combining factor models with dynamic programming for joint segmentation of correlated time-series, including a model selection method and an R package implementation.

Findings

Effective segmentation of correlated time-series demonstrated in simulations.

The method accurately determines the number of breakpoints and factors.

Application to geodesic data shows practical utility.

Abstract

We consider the segmentation of set of correlated time-series, the correlation being allowed to take an arbitrary form but being the same at each time-position. We show that encoding the dependency in a factor model enables us to use the dynamic programming algorithm for the inference of the breakpoints, which remains one the most efficient algorithm. We propose a model selection procedure to determine both the number of breakpoints and the number of factors. This proposed method is implemented in the FASeg R package, which is available on the CRAN. We demonstrate the performances of our procedure through simulation experiments and an application to geodesic data is presented.