Change-Point Detection via Cross-Validation

Last updated: October, 27th 2010 - Back to index

Description

A family of change-point detection procedures via cross-validation was proposed in the following paper, where the goal is to detect changes in the mean of a signal that can be heteroscedastic:
(2010) Sylvain Arlot, Alain Celisse. Segmentation of the mean of heteroscedastic data via cross-validation. Statistics and Computing, 1--20. DOI: 10.1007/s11222-010-9196-x.
[journal]   [pdf.gz]   [supplementary material, pdf.gz]   [short version, 41e Journées de Statistique, SFDS]   [Hal]   [arXiv]

An extensive simulation study was proposed in the above paper, where the new procedures are compared to several others.
The present software allows to generate synthetic data as in the paper, and to perform the new procedures, and some of the previous ones for comparison.
This code is not meant to be fast, but mainly to allow reproducibility of the results presented in the above paper.

Matlab code - version 1.0

The main file in the archive is demo.m. Launch it (and tune its parameter) to compare several procedures considered in the paper on your own data or on simulated data. All other files are subfonctions, see README_CHPTCV for more details.

Change-Point Detection via Cross-Validation - version 1.0


The Change-Point Detection via Cross-Validation package is provided free for non-commercial use under the terms of the GNU General Public License. If you have any questions or comments regarding this package, or if you want to report any bugs, please send an e-mail to sylvain.arlotREMOVETHIS@u-psud.fr. The current version 1.0 has been released on October 27th 2010. Check this web page regularly for newer versions.

THIS SOFTWARE IS PROVIDED BY THE AUTHORS ``AS IS'' AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.