Some Exponential Alternatives to the Normal Model in Time Series Analysis
Abstract
An alternative to the normal model in time series analysis is presented wherein the sequence of random variables have exponential and mixed-exponential marginal distributions. The moving-average and autoregressive models are analyzed with respect to serial correlations and conditional expectations. The mixed-exponential autoregressive model and the mixed autoregressive-moving-average model with exponential marginals are presented. A method of estimating the serial correlation coefficient is examined, and digital computer simulation of several of the models are given.
Document Details
- Document Type
- Technical Report
- Publication Date
- Jun 01, 1975
- Accession Number
- AD1121321
Entities
People
- Eric S. Hodson
Organizations
- Naval Postgraduate School