Fast detection of block boundaries in block-wise constant matrices

Vincent Brault, Julien Chiquet, Céline Lévy-Leduc

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

Abstract

We propose a novel approach for estimating the location of block boundaries (change-points) in a random matrix consisting of a block wise constant matrix observed in white noise. Our method consists in rephrasing this task as a variable selection issue. We use a penalized least-squares criterion with an ℓ1-type penalty for dealing with this problem. We first provide some theoretical results ensuring the consistency of our change-point estimators. Then, we explain how to implement our method in a very efficient way. Finally, we provide some empirical evidence to support our claims and apply our approach to data coming from molecular biology which can be used for better understanding the structure of the chromatin.

Original languageEnglish
Title of host publicationMachine Learning and Data Mining in Pattern Recognition - 12th International Conference, MLDM 2016, Proceedings
EditorsPetra Perner
PublisherSpringer Verlag
Pages214-228
Number of pages15
ISBN (Print)9783319419190
DOIs
Publication statusPublished - 1 Jan 2016
Externally publishedYes
Event12th International Conference on Machine Learning and Data Mining in Pattern Recognition, MLDM 2016 - New York, United States
Duration: 16 Jul 201621 Jul 2016

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume9729
ISSN (Print)0302-9743
ISSN (Electronic)1611-3349

Conference

Conference12th International Conference on Machine Learning and Data Mining in Pattern Recognition, MLDM 2016
Country/TerritoryUnited States
CityNew York
Period16/07/1621/07/16

Keywords

  • Change-points
  • HiC experiments
  • High-dimensional sparse linear model

Fingerprint

Dive into the research topics of 'Fast detection of block boundaries in block-wise constant matrices'. Together they form a unique fingerprint.

Cite this