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

Download:

Current browse context:

stat.AP

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 > Applications

Title: Necessary and Sufficient Condition for Nonsingular Fisher Information Matrix in ARMA Models

Authors: A. Ian McLeod
Abstract: It is demonstrated that a necessary and sufficient condition that the Fisher information matrix of an ARMA model be nonsingular is that the model not be redundant, that is, the autoregressive and moving-average polynomials do not share common roots.
Comments: 5 pages
Subjects: Applications (stat.AP); Statistics Theory (math.ST)
MSC classes: 62M10, 91B84
Journal reference: The American Statistician 53 (1) June 1999
DOI: 10.1080/00031305.1999.10474433
Cite as: arXiv:1611.01387 [stat.AP]
  (or arXiv:1611.01387v1 [stat.AP] for this version)

Submission history

From: Angus Ian McLeod [view email]
[v1] Fri, 4 Nov 2016 14:14:54 GMT (3kb)

Link back to: arXiv, form interface, contact.