We gratefully acknowledge support from
the Simons Foundation and member institutions.
Full-text links:

Download:

Current browse context:

stat.ME

Change to browse by:

References & Citations

Bookmark

(what is this?)
CiteULike logo BibSonomy logo Mendeley logo del.icio.us logo Digg logo Reddit logo

Statistics > Methodology

Title: Multiscale change point detection via gradual bandwidth adjustment in moving sum processes

Abstract: A method for the detection of changes in the expectation in univariate sequences is provided. Moving sum processes are studied. These rely on the selection of a tuning bandwidth. Here, a framework to overcome bandwidth selection is presented - the bandwidth adjusts gradually. For that, moving sum processes are made dependent on both time and the bandwidth: the domain becomes a triangle. On the triangle, paths are constructed which systematically lead to change points. An algorithm is provided that estimates change points by subsequent consideration of paths. Strong consistency for the number and location of change points is shown. Simulations support estimation precision. A companion R-package mscp is made available on CRAN.
Comments: 29 pages, 10 figures
Subjects: Methodology (stat.ME)
MSC classes: 62G20, 62M99
Cite as: arXiv:2103.01060 [stat.ME]
  (or arXiv:2103.01060v2 [stat.ME] for this version)

Submission history

From: Tijana Levajkovic [view email]
[v1] Mon, 1 Mar 2021 15:15:46 GMT (467kb,D)
[v2] Mon, 2 Jan 2023 19:03:46 GMT (496kb,D)

Link back to: arXiv, form interface, contact.