Changing Time Series
Abstract
The statistical analysis of time series data is a challenging and fascinating topic. Much has appeared in the statistical and engineering literature concerning time series analysis, however very few people have attempted the Bayesian approach, but this approach by the principal investigator has been quite successful. Virtually nothing from the Bayesian paradigm had been until the principal investigator began to investigate this area in the early 1980's. The motivation for using the Bayesian method was to eliminate some of the problems left unresolved by the Box-Jenkins approach. The objectives of the proposed research were quite ambitions because the Bayesian methods of analyzing non-changing time series (such as with the ARMA process) had not been tried, thus the first order of business was to try a Bayesian solution on ARMA processes. Later, processes the time changing parameters were analyzed, thus satisfying the objectives of the proposed research.
Document Details
- Document Type
- Technical Report
- Publication Date
- Aug 26, 1986
- Accession Number
- ADA178030
Entities
People
- Lyle Broemeling
Organizations
- Oklahoma State University–Stillwater