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 ScienceWISE logo

Statistics > Methodology

Title: Fast implementation of partial least squares for function-on-function regression

Authors: Zhiyang Zhou
Abstract: People employ the function-on-function regression to model the relationship between two random curves. Fitting this model, widely used strategies include algorithms falling into the framework of functional partial least squares (typically requiring iterative eigen-decomposition). Here we introduce a route of functional partial least squares based upon Krylov subspaces. It can be expressed in two forms equivalent to each other (in exact arithmetic): one is non-iterative with explicit forms of estimators and predictions, facilitating the theoretical derivation and potential extensions (to more complex models); the other one stabilizes numerical outputs. The consistence of estimators and predictions is established under regularity conditions. Our proposal is highlighted as it is less computationally involved. Meanwhile, it is competitive in terms of both estimation and prediction accuracy.
Comments: 25 pages, 2 figures and 2 tables
Subjects: Methodology (stat.ME)
MSC classes: 62G08
Cite as: arXiv:2005.04798 [stat.ME]
  (or arXiv:2005.04798v2 [stat.ME] for this version)

Submission history

From: Zhiyang Zhou [view email]
[v1] Sun, 10 May 2020 22:13:30 GMT (49kb,D)
[v2] Fri, 18 Dec 2020 04:52:42 GMT (105kb,D)

Link back to: arXiv, form interface, contact.