Modern Empirical Statistical Spectral Analysis.
Abstract
This paper has two aims: to provide perspectives on the diverse paths of analysis which are available in 1980 to estimate the spectrum of an observed time series; and to describe proposals for optimal statistical spectral estimation procedures which combine autoregressive spectral estimators and log spectral estimators. It is proposed that empirical statistical spectral analysis should be an adaptive procedure for forming an iterative spectral estimator (an iterative estimator is one composed of estimators obtained in different steps of the analysis). There are three parts: (I). Basic concepts of time series spectral analysis; (II). Entropy distances, autoregressive spectral estimators and log spectral estimators; (III). An outline of empirical spectral analysis. (Author)
Document Details
- Document Type
- Technical Report
- Publication Date
- May 01, 1980
- Accession Number
- ADA084142
Entities
People
- Emanuel Parzen
Organizations
- Texas A&M University