Inference for response-limited time-series models

Research output: Contribution to journalArticle

1 Scopus citations

Abstract

This paper introduces a robust time series estimation procedure based on a new nonlinear time-series model called the response-limited ARMA model. Under this procedure, in contrast to location or regression, the uncorrected pseudo-observations, which produce the parameter estimates when submitted to a standard analysis, also yield satisfactory standard errors for inference purposes.

Original languageEnglish (US)
Pages (from-to)2587-2596
Number of pages10
JournalCommunications in Statistics - Theory and Methods
Volume11
Issue number22
DOIs
StatePublished - Jan 1 1982

Keywords

  • ARMA model
  • robust estimation

ASJC Scopus subject areas

  • Statistics and Probability

Fingerprint Dive into the research topics of 'Inference for response-limited time-series models'. Together they form a unique fingerprint.

  • Cite this