Maximum Entropy Interpretation of Autoregressive Spectral Densities.
Abstract
A new proof is given of the maximum entropy characterization of autoregressive spectral densities as models for the spectral density of a stationary time series. The new proof is presented in parallel with a proof of the maximum entropy characterization of exponential models for probability densities. Concepts of entropy, cross-entropy, and information divergence are defined for probability densities and for spectral densities. (Author)
Document Details
- Document Type
- Technical Report
- Publication Date
- Jun 01, 1982
- Accession Number
- ADA117263
Entities
People
- Emanuel Parzen
Organizations
- Texas A&M University