Estimating the number of change-points in a two-dimensional segmentation model without penalization

Abstract : In computational biology, numerous recent studies have been dedicated to the analysis of the chromatin structure within the cell by two-dimensional segmentation methods. Motivated by this application, we consider the problem of retrieving the diagonal blocks in a matrix of observations. The theoretical properties of the least-squares estimators of both the boundaries and the number of blocks proposed by L\'evy-Leduc et al. [2014] are investigated. More precisely, the contribution of the paper is to establish the consistency of these estimators. A surprising consequence of our results is that, contrary to the onedimensional case, a penalty is not needed for retrieving the true number of diagonal blocks. Finally, the results are illustrated on synthetic data.
Document type :
Preprints, Working Papers, ...
Complete list of metadatas

https://hal-agroparistech.archives-ouvertes.fr/hal-01589417
Contributor : Eva Legras <>
Submitted on : Monday, September 18, 2017 - 3:09:32 PM
Last modification on : Monday, July 15, 2019 - 11:30:04 AM

Links full text

Identifiers

  • HAL Id : hal-01589417, version 1
  • ARXIV : 1506.03198

Citation

V. Brault, M. Delattre, E. Lebarbier, T. Mary-Huard, C. Lévy-Leduc. Estimating the number of change-points in a two-dimensional segmentation model without penalization. 2015. ⟨hal-01589417⟩

Share

Metrics

Record views

89