The Asymptotic Distributions of Autoregressive Coefficients
Abstract
The asymptotic distribution of a finite set of autocorrelations is obtained for a time series from a linear stochastic process. The disturbances (or innovations) are martingale differences with bounded variances and bounded mixed fourth-order moments satisfying a uniform conditional square integrability condition. The conditions are weaker than those used previously for such asymptotic distributions. Specific topics include: autocorrelations, asymptotic distributions, martingale differences, bounded second order moments.
Document Details
- Document Type
- Technical Report
- Publication Date
- Apr 01, 1991
- Accession Number
- ADA238538
Entities
People
- Theodore W. Anderson
Organizations
- Stanford University